WorldWideScience

Sample records for germlings show deep

  1. Alexandrium fundyense cyst viability and germling survival in light vs. dark at a constant low temperature

    Science.gov (United States)

    Vahtera, Emil; Crespo, Bibiana G.; McGillicuddy, Dennis J.; Olli, Kalle; Anderson, Donald M.

    2014-05-01

    Both observations and models suggest that large-scale coastal blooms of Alexandrium fundyense in the Gulf of Maine are seeded by deep-bottom cyst accumulation zones (“seed beds”) where cysts germinate from the sediment surface or the overlying near-bottom nepheloid layers at water depths exceeding 100 m. The germling cells and their vegetative progeny are assumed to be subject to mortality while in complete darkness, as they swim to illuminated surface waters. To test the validity of this assumption we conducted laboratory investigations of cyst viability and the survival of the germling cells and their vegetative progeny during prolonged exposure to darkness at a temperature of 6 °C, simulating the conditions in deep Gulf of Maine waters. We isolated cysts from bottom sediments collected in the Gulf of Maine under low red light and incubated them in 96-well tissue culture-plates in culture medium under a 10:14 h light:dark cycle and under complete darkness. Cyst viability was high, with excystment frequency reaching 90% in the illuminated treatment after 30 days and in the dark treatment after 50 days. Average germination rates were 0.062 and 0.038 d-1 for light and dark treatments, respectively. The dark treatment showed an approximately 2-week time lag in maximum germination rates compared to the light treatment. Survival of germlings was considerably lower in the dark treatment. In the light treatments, 47% of germinated cysts produced germlings that were able to survive for 7 days and produce vegetative progeny, i.e., there were live cells in the well along with an empty cyst at least once during the experiment. In the dark treatments 12% of the cysts produced germlings that were able to survive for the same length of time. When dark treatments are scaled to take into account non-darkness related mortality, approximately 28% of the cysts produced germlings that were able to survive for at least 7 days. Even though cysts are able to germinate in darkness

  2. Chitosan Mediates Germling Adhesion in Magnaporthe oryzae and Is Required for Surface Sensing and Germling Morphogenesis

    Science.gov (United States)

    Geoghegan, Ivey A.; Gurr, Sarah J.

    2016-01-01

    The fungal cell wall not only plays a critical role in maintaining cellular integrity, but also forms the interface between fungi and their environment. The composition of the cell wall can therefore influence the interactions of fungi with their physical and biological environments. Chitin, one of the main polysaccharide components of the wall, can be chemically modified by deacetylation. This reaction is catalyzed by a family of enzymes known as chitin deacetylases (CDAs), and results in the formation of chitosan, a polymer of β1,4-glucosamine. Chitosan has previously been shown to accumulate in the cell wall of infection structures in phytopathogenic fungi. Here, it has long been hypothesized to act as a 'stealth' molecule, necessary for full pathogenesis. In this study, we used the crop pathogen and model organism Magnaporthe oryzae to test this hypothesis. We first confirmed that chitosan localizes to the germ tube and appressorium, then deleted CDA genes on the basis of their elevated transcript levels during appressorium differentiation. Germlings of the deletion strains showed loss of chitin deacetylation, and were compromised in their ability to adhere and form appressoria on artificial hydrophobic surfaces. Surprisingly, the addition of exogenous chitosan fully restored germling adhesion and appressorium development. Despite the lack of appressorium development on artificial surfaces, pathogenicity was unaffected in the mutant strains. Further analyses demonstrated that cuticular waxes are sufficient to over-ride the requirement for chitosan during appressorium development on the plant surface. Thus, chitosan does not have a role as a 'stealth' molecule, but instead mediates the adhesion of germlings to surfaces, thereby allowing the perception of the physical stimuli necessary to promote appressorium development. This study thus reveals a novel role for chitosan in phytopathogenic fungi, and gives further insight into the mechanisms governing

  3. Chitosan Mediates Germling Adhesion in Magnaporthe oryzae and Is Required for Surface Sensing and Germling Morphogenesis.

    Directory of Open Access Journals (Sweden)

    Ivey A Geoghegan

    2016-06-01

    Full Text Available The fungal cell wall not only plays a critical role in maintaining cellular integrity, but also forms the interface between fungi and their environment. The composition of the cell wall can therefore influence the interactions of fungi with their physical and biological environments. Chitin, one of the main polysaccharide components of the wall, can be chemically modified by deacetylation. This reaction is catalyzed by a family of enzymes known as chitin deacetylases (CDAs, and results in the formation of chitosan, a polymer of β1,4-glucosamine. Chitosan has previously been shown to accumulate in the cell wall of infection structures in phytopathogenic fungi. Here, it has long been hypothesized to act as a 'stealth' molecule, necessary for full pathogenesis. In this study, we used the crop pathogen and model organism Magnaporthe oryzae to test this hypothesis. We first confirmed that chitosan localizes to the germ tube and appressorium, then deleted CDA genes on the basis of their elevated transcript levels during appressorium differentiation. Germlings of the deletion strains showed loss of chitin deacetylation, and were compromised in their ability to adhere and form appressoria on artificial hydrophobic surfaces. Surprisingly, the addition of exogenous chitosan fully restored germling adhesion and appressorium development. Despite the lack of appressorium development on artificial surfaces, pathogenicity was unaffected in the mutant strains. Further analyses demonstrated that cuticular waxes are sufficient to over-ride the requirement for chitosan during appressorium development on the plant surface. Thus, chitosan does not have a role as a 'stealth' molecule, but instead mediates the adhesion of germlings to surfaces, thereby allowing the perception of the physical stimuli necessary to promote appressorium development. This study thus reveals a novel role for chitosan in phytopathogenic fungi, and gives further insight into the mechanisms

  4. The nucleation of microtubules in Aspergillus nidulans germlings

    Directory of Open Access Journals (Sweden)

    Cristina de Andrade-Monteiro

    1999-09-01

    Full Text Available Microtubules are filaments composed of dimers of alpha- and beta-tubulins, which have a variety of functions in living cells. In fungi, the spindle pole bodies usually have been considered to be microtubule-organizing centers. We used the antimicrotubule drug Benomyl in block/release experiments to depolymerize and repolymerize microtubules in Aspergillus nidulans germlings to learn more about the microtubule nucleation process in this filamentous fungus. Twenty seconds after release from Benomyl short microtubules were formed from several bright (immunofluorescent dots distributed along the germlings, suggesting that microtubule nucleation is randomly distributed in A. nidulans germlings. Since nuclear movement is dependent on microtubules in A. nidulans we analyzed whether mutants defective in nuclear distribution along the growing hyphae (nud mutants have some obvious microtubule defect. Cytoplasmic, astral and spindle microtubules were present and appeared to be normal in all nud mutants. However, significant changes in the percentage of short versus long mitotic spindles were observed in nud mutants. This suggests that some of the nuclei of nud mutants do not reach the late stage of cell division at normal temperatures.Microtúbulos são filamentos compostos por dímeros das tubulinas a e b e têm uma variedade de funções nas células vivas. Em fungos, os corpúsculos polares dos fusos são geralmente considerados os centros organizadores dos microtúbulos. Com o objetivo de contribuir para uma melhor compreensão dos processos de nucleação dos microtúbulos no fungo filamentoso A. nidulans, nós utilizamos a droga antimicrotúbulo Benomil em experimentos de bloqueio e liberação para depolimerizar e repolimerizar os microtúbulos. Após 20 segundos de reincubação em meio sem Benomil, pequenos microtúbulos foram formados a partir de pontos distribuídos pela célula, sugerindo que os pontos de nucleação de microtúbulos s

  5. Cryptic Plutella species show deep divergence despite the capacity to hybridize.

    Science.gov (United States)

    Perry, Kym D; Baker, Gregory J; Powis, Kevin J; Kent, Joanne K; Ward, Christopher M; Baxter, Simon W

    2018-05-29

    Understanding genomic and phenotypic diversity among cryptic pest taxa has important implications for the management of pests and diseases. The diamondback moth, Plutella xylostella L., has been intensively studied due to its ability to evolve insecticide resistance and status as the world's most destructive pest of brassicaceous crops. The surprise discovery of a cryptic species endemic to Australia, Plutella australiana Landry & Hebert, raised questions regarding the distribution, ecological traits and pest status of the two species, the capacity for gene flow and whether specific management was required. Here, we collected Plutella from wild and cultivated brassicaceous plants from 75 locations throughout Australia and screened 1447 individuals to identify mtDNA lineages and Wolbachia infections. We genotyped genome-wide SNP markers using RADseq in coexisting populations of each species. In addition, we assessed reproductive compatibility in crossing experiments and insecticide susceptibility phenotypes using bioassays. The two Plutella species coexisted on wild brassicas and canola crops, but only 10% of Plutella individuals were P. australiana. This species was not found on commercial Brassica vegetable crops, which are routinely sprayed with insecticides. Bioassays found that P. australiana was 19-306 fold more susceptible to four commonly-used insecticides than P. xylostella. Laboratory crosses revealed that reproductive isolation was incomplete but directionally asymmetric between the species. However, genome-wide nuclear SNPs revealed striking differences in genetic diversity and strong population structure between coexisting wild populations of each species. Nuclear diversity was 1.5-fold higher in P. australiana, yet both species showed limited variation in mtDNA. Infection with a single Wolbachia subgroup B strain was fixed in P. australiana, suggesting that a selective sweep contributed to low mtDNA diversity, while a subgroup A strain infected just 1

  6. Deep neural networks show an equivalent and often superior performance to dermatologists in onychomycosis diagnosis: Automatic construction of onychomycosis datasets by region-based convolutional deep neural network.

    Directory of Open Access Journals (Sweden)

    Seung Seog Han

    Full Text Available Although there have been reports of the successful diagnosis of skin disorders using deep learning, unrealistically large clinical image datasets are required for artificial intelligence (AI training. We created datasets of standardized nail images using a region-based convolutional neural network (R-CNN trained to distinguish the nail from the background. We used R-CNN to generate training datasets of 49,567 images, which we then used to fine-tune the ResNet-152 and VGG-19 models. The validation datasets comprised 100 and 194 images from Inje University (B1 and B2 datasets, respectively, 125 images from Hallym University (C dataset, and 939 images from Seoul National University (D dataset. The AI (ensemble model; ResNet-152 + VGG-19 + feedforward neural networks results showed test sensitivity/specificity/ area under the curve values of (96.0 / 94.7 / 0.98, (82.7 / 96.7 / 0.95, (92.3 / 79.3 / 0.93, (87.7 / 69.3 / 0.82 for the B1, B2, C, and D datasets. With a combination of the B1 and C datasets, the AI Youden index was significantly (p = 0.01 higher than that of 42 dermatologists doing the same assessment manually. For B1+C and B2+ D dataset combinations, almost none of the dermatologists performed as well as the AI. By training with a dataset comprising 49,567 images, we achieved a diagnostic accuracy for onychomycosis using deep learning that was superior to that of most of the dermatologists who participated in this study.

  7. Deep neural networks show an equivalent and often superior performance to dermatologists in onychomycosis diagnosis: Automatic construction of onychomycosis datasets by region-based convolutional deep neural network.

    Science.gov (United States)

    Han, Seung Seog; Park, Gyeong Hun; Lim, Woohyung; Kim, Myoung Shin; Na, Jung Im; Park, Ilwoo; Chang, Sung Eun

    2018-01-01

    Although there have been reports of the successful diagnosis of skin disorders using deep learning, unrealistically large clinical image datasets are required for artificial intelligence (AI) training. We created datasets of standardized nail images using a region-based convolutional neural network (R-CNN) trained to distinguish the nail from the background. We used R-CNN to generate training datasets of 49,567 images, which we then used to fine-tune the ResNet-152 and VGG-19 models. The validation datasets comprised 100 and 194 images from Inje University (B1 and B2 datasets, respectively), 125 images from Hallym University (C dataset), and 939 images from Seoul National University (D dataset). The AI (ensemble model; ResNet-152 + VGG-19 + feedforward neural networks) results showed test sensitivity/specificity/ area under the curve values of (96.0 / 94.7 / 0.98), (82.7 / 96.7 / 0.95), (92.3 / 79.3 / 0.93), (87.7 / 69.3 / 0.82) for the B1, B2, C, and D datasets. With a combination of the B1 and C datasets, the AI Youden index was significantly (p = 0.01) higher than that of 42 dermatologists doing the same assessment manually. For B1+C and B2+ D dataset combinations, almost none of the dermatologists performed as well as the AI. By training with a dataset comprising 49,567 images, we achieved a diagnostic accuracy for onychomycosis using deep learning that was superior to that of most of the dermatologists who participated in this study.

  8. Conquered from the deep sea? A new deep-sea isopod species from the Antarctic shelf shows pattern of recent colonization.

    Directory of Open Access Journals (Sweden)

    Torben Riehl

    Full Text Available The Amundsen Sea, Antarctica, is amongst the most rapidly changing environments of the world. Its benthic inhabitants are barely known and the BIOPEARL 2 project was one of the first to biologically explore this region. Collected during this expedition, Macrostylis roaldi sp. nov. is described as the first isopod discovered on the Amundsen-Sea shelf. Amongst many characteristic features, the most obvious characters unique for M. roaldi are the rather short pleotelson and short operculum as well as the trapezoid shape of the pleotelson in adult males. We used DNA barcodes (COI and additional mitochondrial markers (12S, 16S to reciprocally illuminate morphological results and nucleotide variability. In contrast to many other deep-sea isopods, this species is common and shows a wide distribution. Its range spreads from Pine Island Bay at inner shelf right to the shelf break and across 1,000 m bathymetrically. Its gene pool is homogenized across space and depth. This is indicative for a genetic bottleneck or a recent colonization history. Our results suggest further that migratory or dispersal capabilities of some species of brooding macrobenthos have been underestimated. This might be relevant for the species' potential to cope with effects of climate change. To determine where this species could have survived the last glacial period, alternative refuge possibilities are discussed.

  9. Microscopic and infrared spectroscopic comparison of the underwater adhesives produced by germlings of the brown seaweed species Durvillaea antarctica and Hormosira banksii.

    Science.gov (United States)

    Dimartino, Simone; Savory, David M; Fraser-Miller, Sara J; Gordon, Keith C; McQuillan, A James

    2016-04-01

    Adhesives from marine organisms are often the source of inspiration for the development of glues able to create durable bonds in wet environments. In this work, we investigated the adhesive secretions produced by germlings of two large seaweed species from the South Pacific, Durvillaea antarctica, also named 'the strongest kelp in the word', and its close relative Hormosira banksii The comparative analysis was based on optical and scanning electron microscopy imaging as well as Fourier transform infrared (FTIR) spectroscopy and principal component analysis (PCA). For both species, the egg surface presents peripheral vesicles which are released soon after fertilization to discharge a primary adhesive. This is characterized by peaks representative of carbohydrate molecules. A secondary protein-based adhesive is then secreted in the early developmental stages of the germlings. Energy dispersive X-ray, FTIR and PCA indicate that D. antarctica secretions also contain sulfated moieties, and become cross-linked with time, both conferring strong adhesive and cohesive properties. On the other hand, H. banksii secretions are complemented by the putative adhesive phlorotannins, and are characterized by a simple mechanism in which all constituents are released with the same rate and with no apparent cross-linking. It is also noted that the release of adhesive materials appears to be faster and more copious in D. antarctica than in H. banksii Overall, this study highlights that both quantity and quality of the adhesives matter in explaining the superior attachment ability of D. antarctica. © 2016 The Author(s).

  10. Deep sequencing shows that oocytes are not prone to accumulate mtDNA heteroplasmic mutations during ovarian ageing.

    Science.gov (United States)

    Boucret, L; Bris, C; Seegers, V; Goudenège, D; Desquiret-Dumas, V; Domin-Bernhard, M; Ferré-L'Hotellier, V; Bouet, P E; Descamps, P; Reynier, P; Procaccio, V; May-Panloup, P

    2017-10-01

    Does ovarian ageing increase the number of heteroplasmic mitochondrial DNA (mtDNA) point mutations in oocytes? Our results suggest that oocytes are not subject to the accumulation of mtDNA point mutations during ovarian ageing. Ageing is associated with the alteration of mtDNA integrity in various tissues. Primary oocytes, present in the ovary since embryonic life, may accumulate mtDNA mutations during the process of ovarian ageing. This was an observational study of 53 immature oocyte-cumulus complexes retrieved from 35 women undergoing IVF at the University Hospital of Angers, France, from March 2013 to March 2014. The women were classified in two groups, one including 19 women showing signs of ovarian ageing objectified by a diminished ovarian reserve (DOR), and the other, including 16 women with a normal ovarian reserve (NOR), which served as a control group. mtDNA was extracted from isolated oocytes, and from their corresponding cumulus cells (CCs) considered as a somatic cell compartment. The average mtDNA content of each sample was assessed by using a quantitative real-time PCR technique. Deep sequencing was performed using the Ion Torrent Proton for Next-Generation Sequencing. Signal processing and base calling were done by the embedded pre-processing pipeline and the variants were analyzed using an in-house workflow. The distribution of the different variants between DOR and NOR patients, on one hand, and oocyte and CCs, on the other, was analyzed with the generalized mixed linear model to take into account the cluster of cells belonging to a given mother. There were no significant differences between the numbers of mtDNA variants between the DOR and the NOR patients, either in the oocytes (P = 0.867) or in the surrounding CCs (P = 0.154). There were also no differences in terms of variants with potential functional consequences. De-novo mtDNA variants were found in 28% of the oocytes and in 66% of the CCs with the mean number of variants being

  11. Microbiome analysis shows enrichment for specific bacteria in separate anatomical regions of the deep-sea carnivorous sponge Chondrocladia grandis.

    Science.gov (United States)

    Verhoeven, Joost T P; Kavanagh, Alana N; Dufour, Suzanne C

    2017-01-01

    The Cladorhizidae is a unique family of carnivorous marine sponges characterised by either the absence or reduction of the aquiferous system and by the presence of specialised structures to trap and digest mesoplanktonic prey. Previous studies have postulated a key role of host-associated bacteria in enabling carnivory in this family of sponges. In this study, we employed high-throughput Illumina-based sequencing to identify the bacterial community associated with four individuals of the deep-sea sponge Chondrocladia grandis sampled in the Gulf of Maine. By characterising the V6 through V8 region of the 16S rRNA gene, we compared the bacterial community composition and diversity in three distinct anatomical regions with predicted involvement in prey capture (sphere), support (axis) and benthic substrate attachment (root). A high abundance of Tenacibaculum, a known siderophore producing bacterial genus, was present in all anatomical regions and specimens. The abundance of Colwellia and Roseobacter was greater in sphere and axis samples, and bacteria from the hydrocarbon-degrading Robiginitomaculum genus were most abundant in the root. This first description of the bacterial community associated with C. grandis provides novel insights into the contribution of bacteria to the carnivorous lifestyle while laying foundations for future cladorhizid symbiosis studies. © FEMS 2016. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.

  12. Deep Percolation in Arid Piedmont Slopes: Multiple Lines of Evidence Show How Land Use Change and Ecohydrological Properties Affect Groundwater Recharge

    Science.gov (United States)

    Schreiner-McGraw, A.; Vivoni, E. R.; Browning, D. M.

    2017-12-01

    A critical hydrologic process in arid regions is the contribution of episodic streamflow in ephemeral channels to groundwater recharge. This process has traditionally been studied in channels that drain large watersheds (10s to 100s km2). In this study, we aim to characterize the provision of the ecosystem services of surface and groundwater supply in a first-order watershed (4.6 ha) in an arid piedmont slope of the Jornada Experimental Range (JER). We use an observational and modeling approach to estimate deep percolation. During a 6 year study period, we observed 428 mm of percolation (P) and 39 mm of runoff (Q); ratios of P to rainfall (R) of P/R = 0.27 and Q/R = 0.02. Utilizing an instrument network and site measurements, we determine that percolation occurs primarily inside channel reaches when these receive runoff from upland hillslopes and find that a monthly rainfall threshold of 62 mm is needed for significant percolation to be generated. In order to quantify the mechanisms leading to this threshold response, we develop a channel transmission loss module for the TIN-based Real-time Integrated Basin Simulator (tRIBS) and test the model thoroughly against the available observations over the study period. For these purposes, we make use of image classifications from Unmanned Aerial Vehicle flights, a ground-based phenocam, and species-level measurements to parameterize vegetation processes in the model. We then conduct an extensive set of sensitivity experiments to determine the relative roles of channel, soil, and vegetation properties on modifying the relation between monthly rainfall and percolation. Additionally, we test how the observed vegetation transitions in the JER over the last 150 years affect the deep percolation and runoff estimates. By quantifying mechanisms through which vegetation changes affect water resource provision, this work provides new insights on the ecohydrological controls on the water yield of arid piedmont slopes.

  13. Background synaptic activity in rat entorhinal cortex shows a progressively greater dominance of inhibition over excitation from deep to superficial layers.

    Directory of Open Access Journals (Sweden)

    Stuart David Greenhill

    Full Text Available The entorhinal cortex (EC controls hippocampal input and output, playing major roles in memory and spatial navigation. Different layers of the EC subserve different functions and a number of studies have compared properties of neurones across layers. We have studied synaptic inhibition and excitation in EC neurones, and we have previously compared spontaneous synaptic release of glutamate and GABA using patch clamp recordings of synaptic currents in principal neurones of layers II (L2 and V (L5. Here, we add comparative studies in layer III (L3. Such studies essentially look at neuronal activity from a presynaptic viewpoint. To correlate this with the postsynaptic consequences of spontaneous transmitter release, we have determined global postsynaptic conductances mediated by the two transmitters, using a method to estimate conductances from membrane potential fluctuations. We have previously presented some of this data for L3 and now extend to L2 and L5. Inhibition dominates excitation in all layers but the ratio follows a clear rank order (highest to lowest of L2>L3>L5. The variance of the background conductances was markedly higher for excitation and inhibition in L2 compared to L3 or L5. We also show that induction of synchronized network epileptiform activity by blockade of GABA inhibition reveals a relative reluctance of L2 to participate in such activity. This was associated with maintenance of a dominant background inhibition in L2, whereas in L3 and L5 the absolute level of inhibition fell below that of excitation, coincident with the appearance of synchronized discharges. Further experiments identified potential roles for competition for bicuculline by ambient GABA at the GABAA receptor, and strychnine-sensitive glycine receptors in residual inhibition in L2. We discuss our results in terms of control of excitability in neuronal subpopulations of EC neurones and what these may suggest for their functional roles.

  14. Deep Incremental Boosting

    OpenAIRE

    Mosca, Alan; Magoulas, George D

    2017-01-01

    This paper introduces Deep Incremental Boosting, a new technique derived from AdaBoost, specifically adapted to work with Deep Learning methods, that reduces the required training time and improves generalisation. We draw inspiration from Transfer of Learning approaches to reduce the start-up time to training each incremental Ensemble member. We show a set of experiments that outlines some preliminary results on some common Deep Learning datasets and discuss the potential improvements Deep In...

  15. Deep frying

    NARCIS (Netherlands)

    Koerten, van K.N.

    2016-01-01

    Deep frying is one of the most used methods in the food processing industry. Though practically any food can be fried, French fries are probably the most well-known deep fried products. The popularity of French fries stems from their unique taste and texture, a crispy outside with a mealy soft

  16. Deep learning

    CERN Document Server

    Goodfellow, Ian; Courville, Aaron

    2016-01-01

    Deep learning is a form of machine learning that enables computers to learn from experience and understand the world in terms of a hierarchy of concepts. Because the computer gathers knowledge from experience, there is no need for a human computer operator to formally specify all the knowledge that the computer needs. The hierarchy of concepts allows the computer to learn complicated concepts by building them out of simpler ones; a graph of these hierarchies would be many layers deep. This book introduces a broad range of topics in deep learning. The text offers mathematical and conceptual background, covering relevant concepts in linear algebra, probability theory and information theory, numerical computation, and machine learning. It describes deep learning techniques used by practitioners in industry, including deep feedforward networks, regularization, optimization algorithms, convolutional networks, sequence modeling, and practical methodology; and it surveys such applications as natural language proces...

  17. Deep Learning in Neuroradiology.

    Science.gov (United States)

    Zaharchuk, G; Gong, E; Wintermark, M; Rubin, D; Langlotz, C P

    2018-02-01

    Deep learning is a form of machine learning using a convolutional neural network architecture that shows tremendous promise for imaging applications. It is increasingly being adapted from its original demonstration in computer vision applications to medical imaging. Because of the high volume and wealth of multimodal imaging information acquired in typical studies, neuroradiology is poised to be an early adopter of deep learning. Compelling deep learning research applications have been demonstrated, and their use is likely to grow rapidly. This review article describes the reasons, outlines the basic methods used to train and test deep learning models, and presents a brief overview of current and potential clinical applications with an emphasis on how they are likely to change future neuroradiology practice. Facility with these methods among neuroimaging researchers and clinicians will be important to channel and harness the vast potential of this new method. © 2018 by American Journal of Neuroradiology.

  18. Deep Learning

    DEFF Research Database (Denmark)

    Jensen, Morten Bornø; Bahnsen, Chris Holmberg; Nasrollahi, Kamal

    2018-01-01

    I løbet af de sidste 10 år er kunstige neurale netværk gået fra at være en støvet, udstødt tekno-logi til at spille en hovedrolle i udviklingen af kunstig intelligens. Dette fænomen kaldes deep learning og er inspireret af hjernens opbygning.......I løbet af de sidste 10 år er kunstige neurale netværk gået fra at være en støvet, udstødt tekno-logi til at spille en hovedrolle i udviklingen af kunstig intelligens. Dette fænomen kaldes deep learning og er inspireret af hjernens opbygning....

  19. Deep geothermics

    International Nuclear Information System (INIS)

    Anon.

    1995-01-01

    The hot-dry-rocks located at 3-4 km of depth correspond to low permeable rocks carrying a large amount of heat. The extraction of this heat usually requires artificial hydraulic fracturing of the rock to increase its permeability before water injection. Hot-dry-rocks geothermics or deep geothermics is not today a commercial channel but only a scientific and technological research field. The Soultz-sous-Forets site (Northern Alsace, France) is characterized by a 6 degrees per meter geothermal gradient and is used as a natural laboratory for deep geothermal and geological studies in the framework of a European research program. Two boreholes have been drilled up to 3600 m of depth in the highly-fractured granite massif beneath the site. The aim is to create a deep heat exchanger using only the natural fracturing for water transfer. A consortium of german, french and italian industrial companies (Pfalzwerke, Badenwerk, EdF and Enel) has been created for a more active participation to the pilot phase. (J.S.). 1 fig., 2 photos

  20. Auxiliary Deep Generative Models

    DEFF Research Database (Denmark)

    Maaløe, Lars; Sønderby, Casper Kaae; Sønderby, Søren Kaae

    2016-01-01

    Deep generative models parameterized by neural networks have recently achieved state-of-the-art performance in unsupervised and semi-supervised learning. We extend deep generative models with auxiliary variables which improves the variational approximation. The auxiliary variables leave...... the generative model unchanged but make the variational distribution more expressive. Inspired by the structure of the auxiliary variable we also propose a model with two stochastic layers and skip connections. Our findings suggest that more expressive and properly specified deep generative models converge...... faster with better results. We show state-of-the-art performance within semi-supervised learning on MNIST (0.96%), SVHN (16.61%) and NORB (9.40%) datasets....

  1. Show-Bix &

    DEFF Research Database (Denmark)

    2014-01-01

    The anti-reenactment 'Show-Bix &' consists of 5 dias projectors, a dial phone, quintophonic sound, and interactive elements. A responsive interface will enable the Dias projectors to show copies of original dias slides from the Show-Bix piece ”March på Stedet”, 265 images in total. The copies are...

  2. Deep smarts.

    Science.gov (United States)

    Leonard, Dorothy; Swap, Walter

    2004-09-01

    When a person sizes up a complex situation and rapidly comes to a decision that proves to be not just good but brilliant, you think, "That was smart." After you watch him do this a few times, you realize you're in the presence of something special. It's not raw brainpower, though that helps. It's not emotional intelligence, either, though that, too, is often involved. It's deep smarts. Deep smarts are not philosophical--they're not"wisdom" in that sense, but they're as close to wisdom as business gets. You see them in the manager who understands when and how to move into a new international market, in the executive who knows just what kind of talk to give when her organization is in crisis, in the technician who can track a product failure back to an interaction between independently produced elements. These are people whose knowledge would be hard to purchase on the open market. Their insight is based on know-how more than on know-what; it comprises a system view as well as expertise in individual areas. Because deep smarts are experienced based and often context specific, they can't be produced overnight or readily imported into an organization. It takes years for an individual to develop them--and no time at all for an organization to lose them when a valued veteran walks out the door. They can be taught, however, with the right techniques. Drawing on their forthcoming book Deep Smarts, Dorothy Leonard and Walter Swap say the best way to transfer such expertise to novices--and, on a larger scale, to make individual knowledge institutional--isn't through PowerPoint slides, a Web site of best practices, online training, project reports, or lectures. Rather, the sage needs to teach the neophyte individually how to draw wisdom from experience. Companies have to be willing to dedicate time and effort to such extensive training, but the investment more than pays for itself.

  3. Deep Brain Stimulation for Parkinson's Disease

    Science.gov (United States)

    ... about the BRAIN initiative, see www.nih.gov/science/brain . Show More Show Less Search Disorders SEARCH SEARCH Definition Treatment Prognosis Clinical Trials Organizations Publications Definition Deep ...

  4. Talking with TV shows

    DEFF Research Database (Denmark)

    Sandvik, Kjetil; Laursen, Ditte

    2014-01-01

    User interaction with radio and television programmes is not a new thing. However, with new cross-media production concepts such as X Factor and Voice, this is changing dramatically. The second-screen logic of these productions encourages viewers, along with TV’s traditional one-way communication...... mode, to communicate on interactive (dialogue-enabling) devices such as laptops, smartphones and tablets. Using the TV show Voice as our example, this article shows how the technological and situational set-up of the production invites viewers to engage in new ways of interaction and communication...

  5. Talk Show Science.

    Science.gov (United States)

    Moore, Mitzi Ruth

    1992-01-01

    Proposes having students perform skits in which they play the roles of the science concepts they are trying to understand. Provides the dialog for a skit in which hot and cold gas molecules are interviewed on a talk show to study how these properties affect wind, rain, and other weather phenomena. (MDH)

  6. Obesity in show cats.

    Science.gov (United States)

    Corbee, R J

    2014-12-01

    Obesity is an important disease with a high prevalence in cats. Because obesity is related to several other diseases, it is important to identify the population at risk. Several risk factors for obesity have been described in the literature. A higher incidence of obesity in certain cat breeds has been suggested. The aim of this study was to determine whether obesity occurs more often in certain breeds. The second aim was to relate the increased prevalence of obesity in certain breeds to the official standards of that breed. To this end, 268 cats of 22 different breeds investigated by determining their body condition score (BCS) on a nine-point scale by inspection and palpation, at two different cat shows. Overall, 45.5% of the show cats had a BCS > 5, and 4.5% of the show cats had a BCS > 7. There were significant differences between breeds, which could be related to the breed standards. Most overweight and obese cats were in the neutered group. It warrants firm discussions with breeders and cat show judges to come to different interpretations of the standards in order to prevent overweight conditions in certain breeds from being the standard of beauty. Neutering predisposes for obesity and requires early nutritional intervention to prevent obese conditions. Journal of Animal Physiology and Animal Nutrition © 2014 Blackwell Verlag GmbH.

  7. Honored Teacher Shows Commitment.

    Science.gov (United States)

    Ratte, Kathy

    1987-01-01

    Part of the acceptance speech of the 1985 National Council for the Social Studies Teacher of the Year, this article describes the censorship experience of this honored social studies teacher. The incident involved the showing of a videotape version of the feature film entitled "The Seduction of Joe Tynan." (JDH)

  8. [Deep vein thrombosis prophylaxis.

    Science.gov (United States)

    Sandoval-Chagoya, Gloria Alejandra; Laniado-Laborín, Rafael

    2013-01-01

    Background: despite the proven effectiveness of preventive therapy for deep vein thrombosis, a significant proportion of patients at risk for thromboembolism do not receive prophylaxis during hospitalization. Our objective was to determine the adherence to thrombosis prophylaxis guidelines in a general hospital as a quality control strategy. Methods: a random audit of clinical charts was conducted at the Tijuana General Hospital, Baja California, Mexico, to determine the degree of adherence to deep vein thrombosis prophylaxis guidelines. The instrument used was the Caprini's checklist for thrombosis risk assessment in adult patients. Results: the sample included 300 patient charts; 182 (60.7 %) were surgical patients and 118 were medical patients. Forty six patients (15.3 %) received deep vein thrombosis pharmacologic prophylaxis; 27.1 % of medical patients received deep vein thrombosis prophylaxis versus 8.3 % of surgical patients (p < 0.0001). Conclusions: our results show that adherence to DVT prophylaxis at our hospital is extremely low. Only 15.3 % of our patients at risk received treatment, and even patients with very high risk received treatment in less than 25 % of the cases. We have implemented strategies to increase compliance with clinical guidelines.

  9. The energy show

    International Nuclear Information System (INIS)

    1988-01-01

    The Energy Show is a new look at the problems of world energy, where our supplies come from, now and in the future. The programme looks at how we need energy to maintain our standards of living. Energy supply is shown as the complicated set of problems it is - that Fossil Fuels are both raw materials and energy sources, that some 'alternatives' so readily suggested as practical options are in reality a long way from being effective. (author)

  10. Showing Value (Editorial

    Directory of Open Access Journals (Sweden)

    Denise Koufogiannakis

    2009-06-01

    Full Text Available When Su Cleyle and I first decided to start Evidence Based Library and Information Practice, one of the things we agreed upon immediately was that the journal be open access. We knew that a major obstacle to librarians using the research literature was that they did not have access to the research literature. Although Su and I are both academic librarians who can access a wide variety of library and information literature from our institutions, we belong to a profession where not everyone has equal access to the research in our field. Without such access to our own body of literature, how can we ever hope for practitioners to use research evidence in their decision making? It would have been contradictory to the principles of evidence based library and information practice to do otherwise.One of the specific groups we thought could use such an open access venue for discovering research literature was school librarians. School librarians are often isolated and lacking access to the research literature that may help them prove to stakeholders the importance of their libraries and their role within schools. Certainly, school libraries have been in decline and the use of evidence to show value is needed. As Ken Haycock noted in his 2003 report, The Crisis in Canada’s School Libraries: The Case for Reform and Reinvestment, “Across the country, teacher-librarians are losing their jobs or being reassigned. Collections are becoming depleted owing to budget cuts. Some principals believe that in the age of the Internet and the classroom workstation, the school library is an artifact” (9. Within this context, school librarians are looking to our research literature for evidence of the impact that school library programs have on learning outcomes and student success. They are integrating that evidence into their practice, and reflecting upon what can be improved locally. They are focusing on students and showing the impact of school libraries and

  11. DeepPy: Pythonic deep learning

    DEFF Research Database (Denmark)

    Larsen, Anders Boesen Lindbo

    This technical report introduces DeepPy – a deep learning framework built on top of NumPy with GPU acceleration. DeepPy bridges the gap between highperformance neural networks and the ease of development from Python/NumPy. Users with a background in scientific computing in Python will quickly...... be able to understand and change the DeepPy codebase as it is mainly implemented using high-level NumPy primitives. Moreover, DeepPy supports complex network architectures by letting the user compose mathematical expressions as directed graphs. The latest version is available at http...

  12. Greedy Deep Dictionary Learning

    OpenAIRE

    Tariyal, Snigdha; Majumdar, Angshul; Singh, Richa; Vatsa, Mayank

    2016-01-01

    In this work we propose a new deep learning tool called deep dictionary learning. Multi-level dictionaries are learnt in a greedy fashion, one layer at a time. This requires solving a simple (shallow) dictionary learning problem, the solution to this is well known. We apply the proposed technique on some benchmark deep learning datasets. We compare our results with other deep learning tools like stacked autoencoder and deep belief network; and state of the art supervised dictionary learning t...

  13. Deep Unfolding for Topic Models.

    Science.gov (United States)

    Chien, Jen-Tzung; Lee, Chao-Hsi

    2018-02-01

    Deep unfolding provides an approach to integrate the probabilistic generative models and the deterministic neural networks. Such an approach is benefited by deep representation, easy interpretation, flexible learning and stochastic modeling. This study develops the unsupervised and supervised learning of deep unfolded topic models for document representation and classification. Conventionally, the unsupervised and supervised topic models are inferred via the variational inference algorithm where the model parameters are estimated by maximizing the lower bound of logarithm of marginal likelihood using input documents without and with class labels, respectively. The representation capability or classification accuracy is constrained by the variational lower bound and the tied model parameters across inference procedure. This paper aims to relax these constraints by directly maximizing the end performance criterion and continuously untying the parameters in learning process via deep unfolding inference (DUI). The inference procedure is treated as the layer-wise learning in a deep neural network. The end performance is iteratively improved by using the estimated topic parameters according to the exponentiated updates. Deep learning of topic models is therefore implemented through a back-propagation procedure. Experimental results show the merits of DUI with increasing number of layers compared with variational inference in unsupervised as well as supervised topic models.

  14. Taoism and Deep Ecology.

    Science.gov (United States)

    Sylvan, Richard; Bennett, David

    1988-01-01

    Contrasted are the philosophies of Deep Ecology and ancient Chinese. Discusses the cosmology, morality, lifestyle, views of power, politics, and environmental philosophies of each. Concludes that Deep Ecology could gain much from Taoism. (CW)

  15. Hot, deep origin of petroleum: deep basin evidence and application

    Science.gov (United States)

    Price, Leigh C.

    1978-01-01

    Use of the model of a hot deep origin of oil places rigid constraints on the migration and entrapment of crude oil. Specifically, oil originating from depth migrates vertically up faults and is emplaced in traps at shallower depths. Review of petroleum-producing basins worldwide shows oil occurrence in these basins conforms to the restraints of and therefore supports the hypothesis. Most of the world's oil is found in the very deepest sedimentary basins, and production over or adjacent to the deep basin is cut by or directly updip from faults dipping into the basin deep. Generally the greater the fault throw the greater the reserves. Fault-block highs next to deep sedimentary troughs are the best target areas by the present concept. Traps along major basin-forming faults are quite prospective. The structural style of a basin governs the distribution, types, and amounts of hydrocarbons expected and hence the exploration strategy. Production in delta depocenters (Niger) is in structures cut by or updip from major growth faults, and structures not associated with such faults are barren. Production in block fault basins is on horsts next to deep sedimentary troughs (Sirte, North Sea). In basins whose sediment thickness, structure and geologic history are known to a moderate degree, the main oil occurrences can be specifically predicted by analysis of fault systems and possible hydrocarbon migration routes. Use of the concept permits the identification of significant targets which have either been downgraded or ignored in the past, such as production in or just updip from thrust belts, stratigraphic traps over the deep basin associated with major faulting, production over the basin deep, and regional stratigraphic trapping updip from established production along major fault zones.

  16. Deep Space Telecommunications

    Science.gov (United States)

    Kuiper, T. B. H.; Resch, G. M.

    2000-01-01

    The increasing load on NASA's deep Space Network, the new capabilities for deep space missions inherent in a next-generation radio telescope, and the potential of new telescope technology for reducing construction and operation costs suggest a natural marriage between radio astronomy and deep space telecommunications in developing advanced radio telescope concepts.

  17. Deep learning with Python

    CERN Document Server

    Chollet, Francois

    2018-01-01

    DESCRIPTION Deep learning is applicable to a widening range of artificial intelligence problems, such as image classification, speech recognition, text classification, question answering, text-to-speech, and optical character recognition. Deep Learning with Python is structured around a series of practical code examples that illustrate each new concept introduced and demonstrate best practices. By the time you reach the end of this book, you will have become a Keras expert and will be able to apply deep learning in your own projects. KEY FEATURES • Practical code examples • In-depth introduction to Keras • Teaches the difference between Deep Learning and AI ABOUT THE TECHNOLOGY Deep learning is the technology behind photo tagging systems at Facebook and Google, self-driving cars, speech recognition systems on your smartphone, and much more. AUTHOR BIO Francois Chollet is the author of Keras, one of the most widely used libraries for deep learning in Python. He has been working with deep neural ...

  18. Deep learning evaluation using deep linguistic processing

    OpenAIRE

    Kuhnle, Alexander; Copestake, Ann

    2017-01-01

    We discuss problems with the standard approaches to evaluation for tasks like visual question answering, and argue that artificial data can be used to address these as a complement to current practice. We demonstrate that with the help of existing 'deep' linguistic processing technology we are able to create challenging abstract datasets, which enable us to investigate the language understanding abilities of multimodal deep learning models in detail, as compared to a single performance value ...

  19. Towards deep learning with segregated dendrites.

    Science.gov (United States)

    Guerguiev, Jordan; Lillicrap, Timothy P; Richards, Blake A

    2017-12-05

    Deep learning has led to significant advances in artificial intelligence, in part, by adopting strategies motivated by neurophysiology. However, it is unclear whether deep learning could occur in the real brain. Here, we show that a deep learning algorithm that utilizes multi-compartment neurons might help us to understand how the neocortex optimizes cost functions. Like neocortical pyramidal neurons, neurons in our model receive sensory information and higher-order feedback in electrotonically segregated compartments. Thanks to this segregation, neurons in different layers of the network can coordinate synaptic weight updates. As a result, the network learns to categorize images better than a single layer network. Furthermore, we show that our algorithm takes advantage of multilayer architectures to identify useful higher-order representations-the hallmark of deep learning. This work demonstrates that deep learning can be achieved using segregated dendritic compartments, which may help to explain the morphology of neocortical pyramidal neurons.

  20. Deep learning relevance

    DEFF Research Database (Denmark)

    Lioma, Christina; Larsen, Birger; Petersen, Casper

    2016-01-01

    train a Recurrent Neural Network (RNN) on existing relevant information to that query. We then use the RNN to "deep learn" a single, synthetic, and we assume, relevant document for that query. We design a crowdsourcing experiment to assess how relevant the "deep learned" document is, compared...... to existing relevant documents. Users are shown a query and four wordclouds (of three existing relevant documents and our deep learned synthetic document). The synthetic document is ranked on average most relevant of all....

  1. DAPs: Deep Action Proposals for Action Understanding

    KAUST Repository

    Escorcia, Victor; Caba Heilbron, Fabian; Niebles, Juan Carlos; Ghanem, Bernard

    2016-01-01

    action proposals from long videos. We show how to take advantage of the vast capacity of deep learning models and memory cells to retrieve from untrimmed videos temporal segments, which are likely to contain actions. A comprehensive evaluation indicates

  2. Ploughing the deep sea floor.

    Science.gov (United States)

    Puig, Pere; Canals, Miquel; Company, Joan B; Martín, Jacobo; Amblas, David; Lastras, Galderic; Palanques, Albert

    2012-09-13

    Bottom trawling is a non-selective commercial fishing technique whereby heavy nets and gear are pulled along the sea floor. The direct impact of this technique on fish populations and benthic communities has received much attention, but trawling can also modify the physical properties of seafloor sediments, water–sediment chemical exchanges and sediment fluxes. Most of the studies addressing the physical disturbances of trawl gear on the seabed have been undertaken in coastal and shelf environments, however, where the capacity of trawling to modify the seafloor morphology coexists with high-energy natural processes driving sediment erosion, transport and deposition. Here we show that on upper continental slopes, the reworking of the deep sea floor by trawling gradually modifies the shape of the submarine landscape over large spatial scales. We found that trawling-induced sediment displacement and removal from fishing grounds causes the morphology of the deep sea floor to become smoother over time, reducing its original complexity as shown by high-resolution seafloor relief maps. Our results suggest that in recent decades, following the industrialization of fishing fleets, bottom trawling has become an important driver of deep seascape evolution. Given the global dimension of this type of fishery, we anticipate that the morphology of the upper continental slope in many parts of the world’s oceans could be altered by intensive bottom trawling, producing comparable effects on the deep sea floor to those generated by agricultural ploughing on land.

  3. Deep Vein Thrombosis

    African Journals Online (AJOL)

    OWNER

    Deep Vein Thrombosis: Risk Factors and Prevention in Surgical Patients. Deep Vein ... preventable morbidity and mortality in hospitalized surgical patients. ... the elderly.3,4 It is very rare before the age ... depends on the risk level; therefore an .... but also in the post-operative period. ... is continuing uncertainty regarding.

  4. Stable architectures for deep neural networks

    Science.gov (United States)

    Haber, Eldad; Ruthotto, Lars

    2018-01-01

    Deep neural networks have become invaluable tools for supervised machine learning, e.g. classification of text or images. While often offering superior results over traditional techniques and successfully expressing complicated patterns in data, deep architectures are known to be challenging to design and train such that they generalize well to new data. Critical issues with deep architectures are numerical instabilities in derivative-based learning algorithms commonly called exploding or vanishing gradients. In this paper, we propose new forward propagation techniques inspired by systems of ordinary differential equations (ODE) that overcome this challenge and lead to well-posed learning problems for arbitrarily deep networks. The backbone of our approach is our interpretation of deep learning as a parameter estimation problem of nonlinear dynamical systems. Given this formulation, we analyze stability and well-posedness of deep learning and use this new understanding to develop new network architectures. We relate the exploding and vanishing gradient phenomenon to the stability of the discrete ODE and present several strategies for stabilizing deep learning for very deep networks. While our new architectures restrict the solution space, several numerical experiments show their competitiveness with state-of-the-art networks.

  5. DeepSimulator: a deep simulator for Nanopore sequencing

    KAUST Repository

    Li, Yu

    2017-12-23

    Motivation: Oxford Nanopore sequencing is a rapidly developed sequencing technology in recent years. To keep pace with the explosion of the downstream data analytical tools, a versatile Nanopore sequencing simulator is needed to complement the experimental data as well as to benchmark those newly developed tools. However, all the currently available simulators are based on simple statistics of the produced reads, which have difficulty in capturing the complex nature of the Nanopore sequencing procedure, the main task of which is the generation of raw electrical current signals. Results: Here we propose a deep learning based simulator, DeepSimulator, to mimic the entire pipeline of Nanopore sequencing. Starting from a given reference genome or assembled contigs, we simulate the electrical current signals by a context-dependent deep learning model, followed by a base-calling procedure to yield simulated reads. This workflow mimics the sequencing procedure more naturally. The thorough experiments performed across four species show that the signals generated by our context-dependent model are more similar to the experimentally obtained signals than the ones generated by the official context-independent pore model. In terms of the simulated reads, we provide a parameter interface to users so that they can obtain the reads with different accuracies ranging from 83% to 97%. The reads generated by the default parameter have almost the same properties as the real data. Two case studies demonstrate the application of DeepSimulator to benefit the development of tools in de novo assembly and in low coverage SNP detection. Availability: The software can be accessed freely at: https://github.com/lykaust15/DeepSimulator.

  6. Deep Echo State Network (DeepESN): A Brief Survey

    OpenAIRE

    Gallicchio, Claudio; Micheli, Alessio

    2017-01-01

    The study of deep recurrent neural networks (RNNs) and, in particular, of deep Reservoir Computing (RC) is gaining an increasing research attention in the neural networks community. The recently introduced deep Echo State Network (deepESN) model opened the way to an extremely efficient approach for designing deep neural networks for temporal data. At the same time, the study of deepESNs allowed to shed light on the intrinsic properties of state dynamics developed by hierarchical compositions ...

  7. Extreme Longevity in Proteinaceous Deep-Sea Corals

    Energy Technology Data Exchange (ETDEWEB)

    Roark, E B; Guilderson, T P; Dunbar, R B; Fallon, S J; Mucciarone, D A

    2009-02-09

    Deep-sea corals are found on hard substrates on seamounts and continental margins world-wide at depths of 300 to {approx}3000 meters. Deep-sea coral communities are hotspots of deep ocean biomass and biodiversity, providing critical habitat for fish and invertebrates. Newly applied radiocarbon age date from the deep water proteinaceous corals Gerardia sp. and Leiopathes glaberrima show that radial growth rates are as low as 4 to 35 {micro}m yr{sup -1} and that individual colony longevities are on the order of thousands of years. The management and conservation of deep sea coral communities is challenged by their commercial harvest for the jewelry trade and damage caused by deep water fishing practices. In light of their unusual longevity, a better understanding of deep sea coral ecology and their interrelationships with associated benthic communities is needed to inform coherent international conservation strategies for these important deep-sea ecosystems.

  8. Deep learning in bioinformatics.

    Science.gov (United States)

    Min, Seonwoo; Lee, Byunghan; Yoon, Sungroh

    2017-09-01

    In the era of big data, transformation of biomedical big data into valuable knowledge has been one of the most important challenges in bioinformatics. Deep learning has advanced rapidly since the early 2000s and now demonstrates state-of-the-art performance in various fields. Accordingly, application of deep learning in bioinformatics to gain insight from data has been emphasized in both academia and industry. Here, we review deep learning in bioinformatics, presenting examples of current research. To provide a useful and comprehensive perspective, we categorize research both by the bioinformatics domain (i.e. omics, biomedical imaging, biomedical signal processing) and deep learning architecture (i.e. deep neural networks, convolutional neural networks, recurrent neural networks, emergent architectures) and present brief descriptions of each study. Additionally, we discuss theoretical and practical issues of deep learning in bioinformatics and suggest future research directions. We believe that this review will provide valuable insights and serve as a starting point for researchers to apply deep learning approaches in their bioinformatics studies. © The Author 2016. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com.

  9. Deep subsurface microbial processes

    Science.gov (United States)

    Lovley, D.R.; Chapelle, F.H.

    1995-01-01

    Information on the microbiology of the deep subsurface is necessary in order to understand the factors controlling the rate and extent of the microbially catalyzed redox reactions that influence the geophysical properties of these environments. Furthermore, there is an increasing threat that deep aquifers, an important drinking water resource, may be contaminated by man's activities, and there is a need to predict the extent to which microbial activity may remediate such contamination. Metabolically active microorganisms can be recovered from a diversity of deep subsurface environments. The available evidence suggests that these microorganisms are responsible for catalyzing the oxidation of organic matter coupled to a variety of electron acceptors just as microorganisms do in surface sediments, but at much slower rates. The technical difficulties in aseptically sampling deep subsurface sediments and the fact that microbial processes in laboratory incubations of deep subsurface material often do not mimic in situ processes frequently necessitate that microbial activity in the deep subsurface be inferred through nonmicrobiological analyses of ground water. These approaches include measurements of dissolved H2, which can predict the predominant microbially catalyzed redox reactions in aquifers, as well as geochemical and groundwater flow modeling, which can be used to estimate the rates of microbial processes. Microorganisms recovered from the deep subsurface have the potential to affect the fate of toxic organics and inorganic contaminants in groundwater. Microbial activity also greatly influences 1 the chemistry of many pristine groundwaters and contributes to such phenomena as porosity development in carbonate aquifers, accumulation of undesirably high concentrations of dissolved iron, and production of methane and hydrogen sulfide. Although the last decade has seen a dramatic increase in interest in deep subsurface microbiology, in comparison with the study of

  10. Deep video deblurring

    KAUST Repository

    Su, Shuochen

    2016-11-25

    Motion blur from camera shake is a major problem in videos captured by hand-held devices. Unlike single-image deblurring, video-based approaches can take advantage of the abundant information that exists across neighboring frames. As a result the best performing methods rely on aligning nearby frames. However, aligning images is a computationally expensive and fragile procedure, and methods that aggregate information must therefore be able to identify which regions have been accurately aligned and which have not, a task which requires high level scene understanding. In this work, we introduce a deep learning solution to video deblurring, where a CNN is trained end-to-end to learn how to accumulate information across frames. To train this network, we collected a dataset of real videos recorded with a high framerate camera, which we use to generate synthetic motion blur for supervision. We show that the features learned from this dataset extend to deblurring motion blur that arises due to camera shake in a wide range of videos, and compare the quality of results to a number of other baselines.

  11. Deep inelastic final states

    International Nuclear Information System (INIS)

    Girardi, G.

    1980-11-01

    In these lectures we attempt to describe the final states of deep inelastic scattering as given by QCD. In the first section we shall briefly comment on the parton model and give the main properties of decay functions which are of interest for the study of semi-inclusive leptoproduction. The second section is devoted to the QCD approach to single hadron leptoproduction. First we recall basic facts on QCD log's and derive after that the evolution equations for the fragmentation functions. For this purpose we make a short detour in e + e - annihilation. The rest of the section is a study of the factorization of long distance effects associated with the initial and final states. We then show how when one includes next to leading QCD corrections one induces factorization breaking and describe the double moments useful for testing such effects. The next section contains a review on the QCD jets in the hadronic final state. We begin by introducing the notion of infrared safe variable and defining a few useful examples. Distributions in these variables are studied to first order in QCD, with some comments on the resummation of logs encountered in higher orders. Finally the last section is a 'gaullimaufry' of jet studies

  12. Risk Aversion in Game Shows

    DEFF Research Database (Denmark)

    Andersen, Steffen; Harrison, Glenn W.; Lau, Morten I.

    2008-01-01

    We review the use of behavior from television game shows to infer risk attitudes. These shows provide evidence when contestants are making decisions over very large stakes, and in a replicated, structured way. Inferences are generally confounded by the subjective assessment of skill in some games......, and the dynamic nature of the task in most games. We consider the game shows Card Sharks, Jeopardy!, Lingo, and finally Deal Or No Deal. We provide a detailed case study of the analyses of Deal Or No Deal, since it is suitable for inference about risk attitudes and has attracted considerable attention....

  13. Measuring performance at trade shows

    DEFF Research Database (Denmark)

    Hansen, Kåre

    2004-01-01

    Trade shows is an increasingly important marketing activity to many companies, but current measures of trade show performance do not adequately capture dimensions important to exhibitors. Based on the marketing literature's outcome and behavior-based control system taxonomy, a model is built...... that captures a outcome-based sales dimension and four behavior-based dimensions (i.e. information-gathering, relationship building, image building, and motivation activities). A 16-item instrument is developed for assessing exhibitors perceptions of their trade show performance. The paper presents evidence...

  14. Deep Learning for ECG Classification

    Science.gov (United States)

    Pyakillya, B.; Kazachenko, N.; Mikhailovsky, N.

    2017-10-01

    The importance of ECG classification is very high now due to many current medical applications where this problem can be stated. Currently, there are many machine learning (ML) solutions which can be used for analyzing and classifying ECG data. However, the main disadvantages of these ML results is use of heuristic hand-crafted or engineered features with shallow feature learning architectures. The problem relies in the possibility not to find most appropriate features which will give high classification accuracy in this ECG problem. One of the proposing solution is to use deep learning architectures where first layers of convolutional neurons behave as feature extractors and in the end some fully-connected (FCN) layers are used for making final decision about ECG classes. In this work the deep learning architecture with 1D convolutional layers and FCN layers for ECG classification is presented and some classification results are showed.

  15. Deep Water Survey Data

    Data.gov (United States)

    National Oceanic and Atmospheric Administration, Department of Commerce — The deep water biodiversity surveys explore and describe the biodiversity of the bathy- and bentho-pelagic nekton using Midwater and bottom trawls centered in the...

  16. Deep Space Habitat Project

    Data.gov (United States)

    National Aeronautics and Space Administration — The Deep Space Habitat was closed out at the end of Fiscal Year 2013 (September 30, 2013). Results and select content have been incorporated into the new Exploration...

  17. Deep inelastic lepton scattering

    International Nuclear Information System (INIS)

    Nachtmann, O.

    1977-01-01

    Deep inelastic electron (muon) nucleon and neutrino nucleon scattering as well as electron positron annihilation into hadrons are reviewed from a theoretical point of view. The emphasis is placed on comparisons of quantum chromodynamics with the data. (orig.) [de

  18. Neuromorphic Deep Learning Machines

    OpenAIRE

    Neftci, E; Augustine, C; Paul, S; Detorakis, G

    2017-01-01

    An ongoing challenge in neuromorphic computing is to devise general and computationally efficient models of inference and learning which are compatible with the spatial and temporal constraints of the brain. One increasingly popular and successful approach is to take inspiration from inference and learning algorithms used in deep neural networks. However, the workhorse of deep learning, the gradient descent Back Propagation (BP) rule, often relies on the immediate availability of network-wide...

  19. Pathogenesis of deep endometriosis.

    Science.gov (United States)

    Gordts, Stephan; Koninckx, Philippe; Brosens, Ivo

    2017-12-01

    The pathophysiology of (deep) endometriosis is still unclear. As originally suggested by Cullen, change the definition "deeper than 5 mm" to "adenomyosis externa." With the discovery of the old European literature on uterine bleeding in 5%-10% of the neonates and histologic evidence that the bleeding represents decidual shedding, it is postulated/hypothesized that endometrial stem/progenitor cells, implanted in the pelvic cavity after birth, may be at the origin of adolescent and even the occasionally premenarcheal pelvic endometriosis. Endometriosis in the adolescent is characterized by angiogenic and hemorrhagic peritoneal and ovarian lesions. The development of deep endometriosis at a later age suggests that deep infiltrating endometriosis is a delayed stage of endometriosis. Another hypothesis is that the endometriotic cell has undergone genetic or epigenetic changes and those specific changes determine the development into deep endometriosis. This is compatible with the hereditary aspects, and with the clonality of deep and cystic ovarian endometriosis. It explains the predisposition and an eventual causal effect by dioxin or radiation. Specific genetic/epigenetic changes could explain the various expressions and thus typical, cystic, and deep endometriosis become three different diseases. Subtle lesions are not a disease until epi(genetic) changes occur. A classification should reflect that deep endometriosis is a specific disease. In conclusion the pathophysiology of deep endometriosis remains debated and the mechanisms of disease progression, as well as the role of genetics and epigenetics in the process, still needs to be unraveled. Copyright © 2017 American Society for Reproductive Medicine. Published by Elsevier Inc. All rights reserved.

  20. Deep Learning and Its Applications in Biomedicine.

    Science.gov (United States)

    Cao, Chensi; Liu, Feng; Tan, Hai; Song, Deshou; Shu, Wenjie; Li, Weizhong; Zhou, Yiming; Bo, Xiaochen; Xie, Zhi

    2018-02-01

    Advances in biological and medical technologies have been providing us explosive volumes of biological and physiological data, such as medical images, electroencephalography, genomic and protein sequences. Learning from these data facilitates the understanding of human health and disease. Developed from artificial neural networks, deep learning-based algorithms show great promise in extracting features and learning patterns from complex data. The aim of this paper is to provide an overview of deep learning techniques and some of the state-of-the-art applications in the biomedical field. We first introduce the development of artificial neural network and deep learning. We then describe two main components of deep learning, i.e., deep learning architectures and model optimization. Subsequently, some examples are demonstrated for deep learning applications, including medical image classification, genomic sequence analysis, as well as protein structure classification and prediction. Finally, we offer our perspectives for the future directions in the field of deep learning. Copyright © 2018. Production and hosting by Elsevier B.V.

  1. Overview of deep learning in medical imaging.

    Science.gov (United States)

    Suzuki, Kenji

    2017-09-01

    The use of machine learning (ML) has been increasing rapidly in the medical imaging field, including computer-aided diagnosis (CAD), radiomics, and medical image analysis. Recently, an ML area called deep learning emerged in the computer vision field and became very popular in many fields. It started from an event in late 2012, when a deep-learning approach based on a convolutional neural network (CNN) won an overwhelming victory in the best-known worldwide computer vision competition, ImageNet Classification. Since then, researchers in virtually all fields, including medical imaging, have started actively participating in the explosively growing field of deep learning. In this paper, the area of deep learning in medical imaging is overviewed, including (1) what was changed in machine learning before and after the introduction of deep learning, (2) what is the source of the power of deep learning, (3) two major deep-learning models: a massive-training artificial neural network (MTANN) and a convolutional neural network (CNN), (4) similarities and differences between the two models, and (5) their applications to medical imaging. This review shows that ML with feature input (or feature-based ML) was dominant before the introduction of deep learning, and that the major and essential difference between ML before and after deep learning is the learning of image data directly without object segmentation or feature extraction; thus, it is the source of the power of deep learning, although the depth of the model is an important attribute. The class of ML with image input (or image-based ML) including deep learning has a long history, but recently gained popularity due to the use of the new terminology, deep learning. There are two major models in this class of ML in medical imaging, MTANN and CNN, which have similarities as well as several differences. In our experience, MTANNs were substantially more efficient in their development, had a higher performance, and required a

  2. DeepNAT: Deep convolutional neural network for segmenting neuroanatomy.

    Science.gov (United States)

    Wachinger, Christian; Reuter, Martin; Klein, Tassilo

    2018-04-15

    We introduce DeepNAT, a 3D Deep convolutional neural network for the automatic segmentation of NeuroAnaTomy in T1-weighted magnetic resonance images. DeepNAT is an end-to-end learning-based approach to brain segmentation that jointly learns an abstract feature representation and a multi-class classification. We propose a 3D patch-based approach, where we do not only predict the center voxel of the patch but also neighbors, which is formulated as multi-task learning. To address a class imbalance problem, we arrange two networks hierarchically, where the first one separates foreground from background, and the second one identifies 25 brain structures on the foreground. Since patches lack spatial context, we augment them with coordinates. To this end, we introduce a novel intrinsic parameterization of the brain volume, formed by eigenfunctions of the Laplace-Beltrami operator. As network architecture, we use three convolutional layers with pooling, batch normalization, and non-linearities, followed by fully connected layers with dropout. The final segmentation is inferred from the probabilistic output of the network with a 3D fully connected conditional random field, which ensures label agreement between close voxels. The roughly 2.7million parameters in the network are learned with stochastic gradient descent. Our results show that DeepNAT compares favorably to state-of-the-art methods. Finally, the purely learning-based method may have a high potential for the adaptation to young, old, or diseased brains by fine-tuning the pre-trained network with a small training sample on the target application, where the availability of larger datasets with manual annotations may boost the overall segmentation accuracy in the future. Copyright © 2017 Elsevier Inc. All rights reserved.

  3. Tokyo Motor Show 2003; Tokyo Motor Show 2003

    Energy Technology Data Exchange (ETDEWEB)

    Joly, E.

    2004-01-01

    The text which follows present the different techniques exposed during the 37. Tokyo Motor Show. The report points out the great tendencies of developments of the Japanese automobile industry. The hybrid electric-powered vehicles or those equipped with fuel cells have been highlighted by the Japanese manufacturers which allow considerable budgets in the research of less polluting vehicles. The exposed models, although being all different according to the manufacturer, use always a hybrid system: fuel cell/battery. The manufacturers have stressed too on the intelligent systems for navigation and safety as well as on the design and comfort. (O.M.)

  4. Why & When Deep Learning Works: Looking Inside Deep Learnings

    OpenAIRE

    Ronen, Ronny

    2017-01-01

    The Intel Collaborative Research Institute for Computational Intelligence (ICRI-CI) has been heavily supporting Machine Learning and Deep Learning research from its foundation in 2012. We have asked six leading ICRI-CI Deep Learning researchers to address the challenge of "Why & When Deep Learning works", with the goal of looking inside Deep Learning, providing insights on how deep networks function, and uncovering key observations on their expressiveness, limitations, and potential. The outp...

  5. Pathways to deep decarbonization - 2015 report

    International Nuclear Information System (INIS)

    Ribera, Teresa; Colombier, Michel; Waisman, Henri; Bataille, Chris; Pierfederici, Roberta; Sachs, Jeffrey; Schmidt-Traub, Guido; Williams, Jim; Segafredo, Laura; Hamburg Coplan, Jill; Pharabod, Ivan; Oury, Christian

    2015-12-01

    In September 2015, the Deep Decarbonization Pathways Project published the Executive Summary of the Pathways to Deep Decarbonization: 2015 Synthesis Report. The full 2015 Synthesis Report was launched in Paris on December 3, 2015, at a technical workshop with the Mitigation Action Plans and Scenarios (MAPS) program. The Deep Decarbonization Pathways Project (DDPP) is a collaborative initiative to understand and show how individual countries can transition to a low-carbon economy and how the world can meet the internationally agreed target of limiting the increase in global mean surface temperature to less than 2 degrees Celsius (deg. C). Achieving the 2 deg. C limit will require that global net emissions of greenhouse gases (GHG) approach zero by the second half of the century. In turn, this will require a profound transformation of energy systems by mid-century through steep declines in carbon intensity in all sectors of the economy, a transition we call 'deep decarbonization'

  6. Reality show: um paradoxo nietzschiano

    Directory of Open Access Journals (Sweden)

    Ilana Feldman

    2011-01-01

    Full Text Available

    O fenômeno dos reality shows - e a subseqüente relação entre imagem e verdade - assenta-se sobre uma série de paradoxos. Tais paradoxos podem ser compreendidos à luz do pensamento do filósofo alemão Friedrich Nietzsche, que, através dos usos de formulações paradoxais, concebia a realidade como um mundo de pura aparência e a verdade como um acréscimo ficcional, como um efeito. A ficção é então tomada, na filosofia de Nietzsche, não em seu aspecto falsificante e desrealizador - como sempre pleiteou nossa tradição metafísica -, mas como condição necessária para que certa espécie de invenção possa operar como verdade. Sendo assim, a própria expressão reality show, através de sua formulação paradoxal, engendra explicitamente um mundo de pura aparência, em que a verdade, a parte reality da proposição, é da ordem do suplemento, daquilo que se acrescenta ficcionalmente - como um adjetivo - a show. O ornamento, nesse caso, passa a ocupar o lugar central, apontando para o efeito produzido: o efeito-de-verdade. Seguindo, então, o pensamento nietzschiano e sua atualização na contemporaneidade, investigaremos de que forma os televisivos “shows de realidade” operam paradoxalmente, em consonância com nossas paradoxais práticas culturais.

  7. DeepSurv: personalized treatment recommender system using a Cox proportional hazards deep neural network.

    Science.gov (United States)

    Katzman, Jared L; Shaham, Uri; Cloninger, Alexander; Bates, Jonathan; Jiang, Tingting; Kluger, Yuval

    2018-02-26

    Medical practitioners use survival models to explore and understand the relationships between patients' covariates (e.g. clinical and genetic features) and the effectiveness of various treatment options. Standard survival models like the linear Cox proportional hazards model require extensive feature engineering or prior medical knowledge to model treatment interaction at an individual level. While nonlinear survival methods, such as neural networks and survival forests, can inherently model these high-level interaction terms, they have yet to be shown as effective treatment recommender systems. We introduce DeepSurv, a Cox proportional hazards deep neural network and state-of-the-art survival method for modeling interactions between a patient's covariates and treatment effectiveness in order to provide personalized treatment recommendations. We perform a number of experiments training DeepSurv on simulated and real survival data. We demonstrate that DeepSurv performs as well as or better than other state-of-the-art survival models and validate that DeepSurv successfully models increasingly complex relationships between a patient's covariates and their risk of failure. We then show how DeepSurv models the relationship between a patient's features and effectiveness of different treatment options to show how DeepSurv can be used to provide individual treatment recommendations. Finally, we train DeepSurv on real clinical studies to demonstrate how it's personalized treatment recommendations would increase the survival time of a set of patients. The predictive and modeling capabilities of DeepSurv will enable medical researchers to use deep neural networks as a tool in their exploration, understanding, and prediction of the effects of a patient's characteristics on their risk of failure.

  8. Deep Learning from Crowds

    DEFF Research Database (Denmark)

    Rodrigues, Filipe; Pereira, Francisco Camara

    Over the last few years, deep learning has revolutionized the field of machine learning by dramatically improving the stateof-the-art in various domains. However, as the size of supervised artificial neural networks grows, typically so does the need for larger labeled datasets. Recently...... networks from crowds. We begin by describing an EM algorithm for jointly learning the parameters of the network and the reliabilities of the annotators. Then, a novel general-purpose crowd layer is proposed, which allows us to train deep neural networks end-to-end, directly from the noisy labels......, crowdsourcing has established itself as an efficient and cost-effective solution for labeling large sets of data in a scalable manner, but it often requires aggregating labels from multiple noisy contributors with different levels of expertise. In this paper, we address the problem of learning deep neural...

  9. Deep boreholes; Tiefe Bohrloecher

    Energy Technology Data Exchange (ETDEWEB)

    Bracke, Guido [Gesellschaft fuer Anlagen- und Reaktorsicherheit gGmbH Koeln (Germany); Charlier, Frank [NSE international nuclear safety engineering gmbh, Aachen (Germany); Geckeis, Horst [Karlsruher Institut fuer Technologie (Germany). Inst. fuer Nukleare Entsorgung; and others

    2016-02-15

    The report on deep boreholes covers the following subject areas: methods for safe enclosure of radioactive wastes, requirements concerning the geological conditions of possible boreholes, reversibility of decisions and retrievability, status of drilling technology. The introduction covers national and international activities. Further chapters deal with the following issues: basic concept of the storage in deep bore holes, status of the drilling technology, safe enclosure, geomechanics and stability, reversibility of decisions, risk scenarios, compliancy with safe4ty requirements and site selection criteria, research and development demand.

  10. Deep Water Acoustics

    Science.gov (United States)

    2016-06-28

    the Deep Water project and participate in the NPAL Workshops, including Art Baggeroer (MIT), J. Beron- Vera (UMiami), M. Brown (UMiami), T...Kathleen E . Wage. The North Pacific Acoustic Laboratory deep-water acoustic propagation experiments in the Philippine Sea. J. Acoust. Soc. Am., 134(4...estimate of the angle α during PhilSea09, made from ADCP measurements at the site of the DVLA. Sim. A B1 B2 B3 C D E F Prof. # 0 4 4 4 5 10 16 20 α

  11. Deep water recycling through time.

    Science.gov (United States)

    Magni, Valentina; Bouilhol, Pierre; van Hunen, Jeroen

    2014-11-01

    We investigate the dehydration processes in subduction zones and their implications for the water cycle throughout Earth's history. We use a numerical tool that combines thermo-mechanical models with a thermodynamic database to examine slab dehydration for present-day and early Earth settings and its consequences for the deep water recycling. We investigate the reactions responsible for releasing water from the crust and the hydrated lithospheric mantle and how they change with subduction velocity ( v s ), slab age ( a ) and mantle temperature (T m ). Our results show that faster slabs dehydrate over a wide area: they start dehydrating shallower and they carry water deeper into the mantle. We parameterize the amount of water that can be carried deep into the mantle, W (×10 5 kg/m 2 ), as a function of v s (cm/yr), a (Myrs), and T m (°C):[Formula: see text]. We generally observe that a 1) 100°C increase in the mantle temperature, or 2) ∼15 Myr decrease of plate age, or 3) decrease in subduction velocity of ∼2 cm/yr all have the same effect on the amount of water retained in the slab at depth, corresponding to a decrease of ∼2.2×10 5 kg/m 2 of H 2 O. We estimate that for present-day conditions ∼26% of the global influx water, or 7×10 8 Tg/Myr of H 2 O, is recycled into the mantle. Using a realistic distribution of subduction parameters, we illustrate that deep water recycling might still be possible in early Earth conditions, although its efficiency would generally decrease. Indeed, 0.5-3.7 × 10 8 Tg/Myr of H 2 O could still be recycled in the mantle at 2.8 Ga. Deep water recycling might be possible even in early Earth conditions We provide a scaling law to estimate the amount of H 2 O flux deep into the mantle Subduction velocity has a a major control on the crustal dehydration pattern.

  12. Contemporary deep recurrent learning for recognition

    Science.gov (United States)

    Iftekharuddin, K. M.; Alam, M.; Vidyaratne, L.

    2017-05-01

    Large-scale feed-forward neural networks have seen intense application in many computer vision problems. However, these networks can get hefty and computationally intensive with increasing complexity of the task. Our work, for the first time in literature, introduces a Cellular Simultaneous Recurrent Network (CSRN) based hierarchical neural network for object detection. CSRN has shown to be more effective to solving complex tasks such as maze traversal and image processing when compared to generic feed forward networks. While deep neural networks (DNN) have exhibited excellent performance in object detection and recognition, such hierarchical structure has largely been absent in neural networks with recurrency. Further, our work introduces deep hierarchy in SRN for object recognition. The simultaneous recurrency results in an unfolding effect of the SRN through time, potentially enabling the design of an arbitrarily deep network. This paper shows experiments using face, facial expression and character recognition tasks using novel deep recurrent model and compares recognition performance with that of generic deep feed forward model. Finally, we demonstrate the flexibility of incorporating our proposed deep SRN based recognition framework in a humanoid robotic platform called NAO.

  13. Analytical estimation show low depth-independent water loss due to vapor flux from deep aquifers

    Science.gov (United States)

    Selker, John S.

    2017-06-01

    Recent articles have provided estimates of evaporative flux from water tables in deserts that span 5 orders of magnitude. In this paper, we present an analytical calculation that indicates aquifer vapor flux to be limited to 0.01 mm/yr for sites where there is negligible recharge and the water table is well over 20 m below the surface. This value arises from the geothermal gradient, and therefore, is nearly independent of the actual depth of the aquifer. The value is in agreement with several numerical studies, but is 500 times lower than recently reported experimental values, and 100 times larger than an earlier analytical estimate.

  14. Colwellia psychrerythraea strains from distant deep sea basins show adaptation to local conditions

    Directory of Open Access Journals (Sweden)

    Stephen M Techtmann

    2016-05-01

    Full Text Available Many studies have shown that microbes, which share nearly identical 16S rRNA genes, can have highly divergent genomes. Microbes from distinct parts of the ocean also exhibit biogeographic patterning. Here we seek to better understand how certain microbes from the same species have adapted for growth under local conditions. The phenotypic and genomic heterogeneity of three strains of Colwellia psychrerythraea was investigated in order to understand adaptions to local environments. Colwellia are psychrophilic heterotrophic marine bacteria ubiquitous in cold marine ecosystems. We have recently isolated two Colwellia strains: ND2E from the Eastern Mediterranean and GAB14E from the Great Australian Bight. The 16S rRNA sequence of these two strains were greater than 98.2% identical to the well-characterized C. psychrerythraea 34H, which was isolated from arctic sediments. Salt tolerance, and carbon source utilization profiles for these strains were determined using Biolog Phenotype Microarrays’. These strains exhibited distinct salt tolerance, which was not associated with the salinity of sites of isolation. The carbon source utilization profiles were distinct with less than half of the tested carbon sources being metabolized by all three strains. Whole genome sequencing revealed that the genomes of these three strains were quite diverse with some genomes having up to 1600 strain-specific genes. Many genes involved in degrading strain-specific carbon sources were identified. There appears to be a link between carbon source utilization and location of isolation with distinctions observed between the Colwellia isolate recovered from sediment compared to water column isolates.

  15. Deep diode atomic battery

    International Nuclear Information System (INIS)

    Anthony, T.R.; Cline, H.E.

    1977-01-01

    A deep diode atomic battery is made from a bulk semiconductor crystal containing three-dimensional arrays of columnar and lamellar P-N junctions. The battery is powered by gamma rays and x-ray emission from a radioactive source embedded in the interior of the semiconductor crystal

  16. Deep Learning Policy Quantization

    NARCIS (Netherlands)

    van de Wolfshaar, Jos; Wiering, Marco; Schomaker, Lambertus

    2018-01-01

    We introduce a novel type of actor-critic approach for deep reinforcement learning which is based on learning vector quantization. We replace the softmax operator of the policy with a more general and more flexible operator that is similar to the robust soft learning vector quantization algorithm.

  17. Deep-sea fungi

    Digital Repository Service at National Institute of Oceanography (India)

    Raghukumar, C; Damare, S.R.

    significant in terms of carbon sequestration (5, 8). In light of this, the diversity, abundance, and role of fungi in deep-sea sediments may form an important link in the global C biogeochemistry. This review focuses on issues related to collection...

  18. Deep inelastic scattering

    International Nuclear Information System (INIS)

    Aubert, J.J.

    1982-01-01

    Deep inelastic lepton-nucleon interaction experiments are renewed. Singlet and non-singlet structure functions are measured and the consistency of the different results is checked. A detailed analysis of the scaling violation is performed in terms of the quantum chromodynamics predictions [fr

  19. Deep Vein Thrombosis

    Centers for Disease Control (CDC) Podcasts

    2012-04-05

    This podcast discusses the risk for deep vein thrombosis in long-distance travelers and ways to minimize that risk.  Created: 4/5/2012 by National Center for Emerging and Zoonotic Infectious Diseases (NCEZID).   Date Released: 4/5/2012.

  20. Deep Learning Microscopy

    KAUST Repository

    Rivenson, Yair; Gorocs, Zoltan; Gunaydin, Harun; Zhang, Yibo; Wang, Hongda; Ozcan, Aydogan

    2017-01-01

    regular optical microscope, without any changes to its design. We blindly tested this deep learning approach using various tissue samples that are imaged with low-resolution and wide-field systems, where the network rapidly outputs an image with remarkably

  1. The deep universe

    CERN Document Server

    Sandage, AR; Longair, MS

    1995-01-01

    Discusses the concept of the deep universe from two conflicting theoretical viewpoints: firstly as a theory embracing the evolution of the universe from the Big Bang to the present; and secondly through observations gleaned over the years on stars, galaxies and clusters.

  2. Teaching for Deep Learning

    Science.gov (United States)

    Smith, Tracy Wilson; Colby, Susan A.

    2007-01-01

    The authors have been engaged in research focused on students' depth of learning as well as teachers' efforts to foster deep learning. Findings from a study examining the teaching practices and student learning outcomes of sixty-four teachers in seventeen different states (Smith et al. 2005) indicated that most of the learning in these classrooms…

  3. Deep Trawl Dataset

    Data.gov (United States)

    National Oceanic and Atmospheric Administration, Department of Commerce — Otter trawl (36' Yankee and 4-seam net deepwater gear) catches from mid-Atlantic slope and canyons at 200 - 800 m depth. Deep-sea (200-800 m depth) flat otter trawls...

  4. Parallel Distributed Processing theory in the age of deep networks

    OpenAIRE

    Bowers, Jeffrey

    2017-01-01

    Parallel Distributed Processing (PDP) models in psychology are the precursors of deep networks used in computer science. However, only PDP models are associated with two core psychological claims, namely, that all knowledge is coded in a distributed format, and cognition is mediated by non-symbolic computations. These claims have long been debated within cognitive science, and recent work with deep networks speaks to this debate. Specifically, single-unit recordings show that deep networks le...

  5. Deep-seated sarcomas of the penis

    Directory of Open Access Journals (Sweden)

    Alberto A. Antunes

    2005-06-01

    Full Text Available Mesenchymal neoplasias represent 5% of tumors affecting the penis. Due to the rarity of such tumors, there is no agreement concerning the best method for staging and managing these patients. Sarcomas of the penis can be classified as deep-seated if they derive from the structures forming the spongy body and the cavernous bodies. Superficial lesions are usually low-grade and show a small tendency towards distant metastasis. In contrast, deep-seated lesions usually show behavior that is more aggressive and have poorer prognosis. The authors report 3 cases of deep-seated primary sarcomas of the penis and review the literature on this rare and aggressive neoplasia.

  6. Deep learning in jet reconstruction at CMS

    CERN Document Server

    Stoye, Markus

    2017-01-01

    Deep learning has led to several breakthroughs outside the field of high energy physics, yet in jet reconstruction for the CMS experiment at the CERN LHC it has not been used so far. This report shows results of applying deep learning strategies to jet reconstruction at the stage of identifying the original parton association of the jet (jet tagging), which is crucial for physics analyses at the LHC experiments. We introduce a custom deep neural network architecture for jet tagging. We compare the performance of this novel method with the other established approaches at CMS and show that the proposed strategy provides a significant improvement. The strategy provides the first multi-class classifier, instead of the few binary classifiers that previously were used, and thus yields more information and in a more convenient way. The performance results obtained with simulation imply a significant improvement for a large number of important physics analysis at the CMS experiment.

  7. DeepMitosis: Mitosis detection via deep detection, verification and segmentation networks.

    Science.gov (United States)

    Li, Chao; Wang, Xinggang; Liu, Wenyu; Latecki, Longin Jan

    2018-04-01

    Mitotic count is a critical predictor of tumor aggressiveness in the breast cancer diagnosis. Nowadays mitosis counting is mainly performed by pathologists manually, which is extremely arduous and time-consuming. In this paper, we propose an accurate method for detecting the mitotic cells from histopathological slides using a novel multi-stage deep learning framework. Our method consists of a deep segmentation network for generating mitosis region when only a weak label is given (i.e., only the centroid pixel of mitosis is annotated), an elaborately designed deep detection network for localizing mitosis by using contextual region information, and a deep verification network for improving detection accuracy by removing false positives. We validate the proposed deep learning method on two widely used Mitosis Detection in Breast Cancer Histological Images (MITOSIS) datasets. Experimental results show that we can achieve the highest F-score on the MITOSIS dataset from ICPR 2012 grand challenge merely using the deep detection network. For the ICPR 2014 MITOSIS dataset that only provides the centroid location of mitosis, we employ the segmentation model to estimate the bounding box annotation for training the deep detection network. We also apply the verification model to eliminate some false positives produced from the detection model. By fusing scores of the detection and verification models, we achieve the state-of-the-art results. Moreover, our method is very fast with GPU computing, which makes it feasible for clinical practice. Copyright © 2018 Elsevier B.V. All rights reserved.

  8. Deep inelastic scattering

    International Nuclear Information System (INIS)

    Zakharov, V.I.

    1977-01-01

    The present status of the quark-parton-gluon picture of deep inelastic scattering is reviewed. The general framework is mostly theoretical and covers investigations since 1970. Predictions of the parton model and of the asymptotically free field theories are compared with experimental data available. The valence quark approximation is concluded to be valid in most cases, but fails to account for the data on the total momentum transfer. On the basis of gluon corrections introduced to the parton model certain predictions concerning both the deep inelastic structure functions and form factors are made. The contributions of gluon exchanges and gluon bremsstrahlung are highlighted. Asymptotic freedom is concluded to be very attractive and provide qualitative explanation to some experimental observations (scaling violations, breaking of the Drell-Yan-West type relations). Lepton-nuclear scattering is pointed out to be helpful in probing the nature of nuclear forces and studying the space-time picture of the parton model

  9. Deep Energy Retrofit

    DEFF Research Database (Denmark)

    Zhivov, Alexander; Lohse, Rüdiger; Rose, Jørgen

    Deep Energy Retrofit – A Guide to Achieving Significant Energy User Reduction with Major Renovation Projects contains recommendations for characteristics of some of core technologies and measures that are based on studies conducted by national teams associated with the International Energy Agency...... Energy Conservation in Buildings and Communities Program (IEA-EBC) Annex 61 (Lohse et al. 2016, Case, et al. 2016, Rose et al. 2016, Yao, et al. 2016, Dake 2014, Stankevica et al. 2016, Kiatreungwattana 2014). Results of these studies provided a base for setting minimum requirements to the building...... envelope-related technologies to make Deep Energy Retrofit feasible and, in many situations, cost effective. Use of energy efficiency measures (EEMs) in addition to core technologies bundle and high-efficiency appliances will foster further energy use reduction. This Guide also provides best practice...

  10. Deep groundwater chemistry

    International Nuclear Information System (INIS)

    Wikberg, P.; Axelsen, K.; Fredlund, F.

    1987-06-01

    Starting in 1977 and up till now a number of places in Sweden have been investigated in order to collect the necessary geological, hydrogeological and chemical data needed for safety analyses of repositories in deep bedrock systems. Only crystalline rock is considered and in many cases this has been gneisses of sedimentary origin but granites and gabbros are also represented. Core drilled holes have been made at nine sites. Up to 15 holes may be core drilled at one site, the deepest down to 1000 m. In addition to this a number of boreholes are percussion drilled at each site to depths of about 100 m. When possible drilling water is taken from percussion drilled holes. The first objective is to survey the hydraulic conditions. Core drilled boreholes and sections selected for sampling of deep groundwater are summarized. (orig./HP)

  11. Deep Reinforcement Fuzzing

    OpenAIRE

    Böttinger, Konstantin; Godefroid, Patrice; Singh, Rishabh

    2018-01-01

    Fuzzing is the process of finding security vulnerabilities in input-processing code by repeatedly testing the code with modified inputs. In this paper, we formalize fuzzing as a reinforcement learning problem using the concept of Markov decision processes. This in turn allows us to apply state-of-the-art deep Q-learning algorithms that optimize rewards, which we define from runtime properties of the program under test. By observing the rewards caused by mutating with a specific set of actions...

  12. Parallel Distributed Processing Theory in the Age of Deep Networks.

    Science.gov (United States)

    Bowers, Jeffrey S

    2017-12-01

    Parallel distributed processing (PDP) models in psychology are the precursors of deep networks used in computer science. However, only PDP models are associated with two core psychological claims, namely that all knowledge is coded in a distributed format and cognition is mediated by non-symbolic computations. These claims have long been debated in cognitive science, and recent work with deep networks speaks to this debate. Specifically, single-unit recordings show that deep networks learn units that respond selectively to meaningful categories, and researchers are finding that deep networks need to be supplemented with symbolic systems to perform some tasks. Given the close links between PDP and deep networks, it is surprising that research with deep networks is challenging PDP theory. Copyright © 2017. Published by Elsevier Ltd.

  13. Deep Visual Attention Prediction

    Science.gov (United States)

    Wang, Wenguan; Shen, Jianbing

    2018-05-01

    In this work, we aim to predict human eye fixation with view-free scenes based on an end-to-end deep learning architecture. Although Convolutional Neural Networks (CNNs) have made substantial improvement on human attention prediction, it is still needed to improve CNN based attention models by efficiently leveraging multi-scale features. Our visual attention network is proposed to capture hierarchical saliency information from deep, coarse layers with global saliency information to shallow, fine layers with local saliency response. Our model is based on a skip-layer network structure, which predicts human attention from multiple convolutional layers with various reception fields. Final saliency prediction is achieved via the cooperation of those global and local predictions. Our model is learned in a deep supervision manner, where supervision is directly fed into multi-level layers, instead of previous approaches of providing supervision only at the output layer and propagating this supervision back to earlier layers. Our model thus incorporates multi-level saliency predictions within a single network, which significantly decreases the redundancy of previous approaches of learning multiple network streams with different input scales. Extensive experimental analysis on various challenging benchmark datasets demonstrate our method yields state-of-the-art performance with competitive inference time.

  14. Top Tagging by Deep Learning Algorithm

    CERN Document Server

    Akil, Ali

    2015-01-01

    In this report I will show the application of a deep learning algorithm on a Monte Carlo simulation sample to test its performance in tagging hadronic decays of boosted top quarks and compare what we get with the results of the application of some other algorithms.

  15. Using Cooperative Structures to Promote Deep Learning

    Science.gov (United States)

    Millis, Barbara J.

    2014-01-01

    The author explores concrete ways to help students learn more and have fun doing it while they support each other's learning. The article specifically shows the relationships between cooperative learning and deep learning. Readers will become familiar with the tenets of cooperative learning and its power to enhance learning--even more so when…

  16. Quantitative phase microscopy using deep neural networks

    Science.gov (United States)

    Li, Shuai; Sinha, Ayan; Lee, Justin; Barbastathis, George

    2018-02-01

    Deep learning has been proven to achieve ground-breaking accuracy in various tasks. In this paper, we implemented a deep neural network (DNN) to achieve phase retrieval in a wide-field microscope. Our DNN utilized the residual neural network (ResNet) architecture and was trained using the data generated by a phase SLM. The results showed that our DNN was able to reconstruct the profile of the phase target qualitatively. In the meantime, large error still existed, which indicated that our approach still need to be improved.

  17. Deep Learning and Developmental Learning: Emergence of Fine-to-Coarse Conceptual Categories at Layers of Deep Belief Network.

    Science.gov (United States)

    Sadeghi, Zahra

    2016-09-01

    In this paper, I investigate conceptual categories derived from developmental processing in a deep neural network. The similarity matrices of deep representation at each layer of neural network are computed and compared with their raw representation. While the clusters generated by raw representation stand at the basic level of abstraction, conceptual categories obtained from deep representation shows a bottom-up transition procedure. Results demonstrate a developmental course of learning from specific to general level of abstraction through learned layers of representations in a deep belief network. © The Author(s) 2016.

  18. Deep learning for SAR image formation

    Science.gov (United States)

    Mason, Eric; Yonel, Bariscan; Yazici, Birsen

    2017-04-01

    The recent success of deep learning has lead to growing interest in applying these methods to signal processing problems. This paper explores the applications of deep learning to synthetic aperture radar (SAR) image formation. We review deep learning from a perspective relevant to SAR image formation. Our objective is to address SAR image formation in the presence of uncertainties in the SAR forward model. We present a recurrent auto-encoder network architecture based on the iterative shrinkage thresholding algorithm (ISTA) that incorporates SAR modeling. We then present an off-line training method using stochastic gradient descent and discuss the challenges and key steps of learning. Lastly, we show experimentally that our method can be used to form focused images in the presence of phase uncertainties. We demonstrate that the resulting algorithm has faster convergence and decreased reconstruction error than that of ISTA.

  19. Deep Red (Profondo Rosso)

    CERN Multimedia

    Cine Club

    2015-01-01

    Wednesday 29 April 2015 at 20:00 CERN Council Chamber    Deep Red (Profondo Rosso) Directed by Dario Argento (Italy, 1975) 126 minutes A psychic who can read minds picks up the thoughts of a murderer in the audience and soon becomes a victim. An English pianist gets involved in solving the murders, but finds many of his avenues of inquiry cut off by new murders, and he begins to wonder how the murderer can track his movements so closely. Original version Italian; English subtitles

  20. Reversible deep disposal

    International Nuclear Information System (INIS)

    2009-10-01

    This presentation, given by the national agency of radioactive waste management (ANDRA) at the meeting of October 8, 2009 of the high committee for the nuclear safety transparency and information (HCTISN), describes the concept of deep reversible disposal for high level/long living radioactive wastes, as considered by the ANDRA in the framework of the program law of June 28, 2006 about the sustainable management of radioactive materials and wastes. The document presents the social and political reasons of reversibility, the technical means considered (containers, disposal cavities, monitoring system, test facilities and industrial prototypes), the decisional process (progressive development and blocked off of the facility, public information and debate). (J.S.)

  1. Deep inelastic neutron scattering

    International Nuclear Information System (INIS)

    Mayers, J.

    1989-03-01

    The report is based on an invited talk given at a conference on ''Neutron Scattering at ISIS: Recent Highlights in Condensed Matter Research'', which was held in Rome, 1988, and is intended as an introduction to the techniques of Deep Inelastic Neutron Scattering. The subject is discussed under the following topic headings:- the impulse approximation I.A., scaling behaviour, kinematical consequences of energy and momentum conservation, examples of measurements, derivation of the I.A., the I.A. in a harmonic system, and validity of the I.A. in neutron scattering. (U.K.)

  2. [Deep mycoses rarely described].

    Science.gov (United States)

    Charles, D

    1986-01-01

    Beside deep mycoses very well known: histoplasmosis, candidosis, cryptococcosis, there are other mycoses less frequently described. Some of them are endemic in some countries: South American blastomycosis in Brazil, coccidioidomycosis in California; some others are cosmopolitan and may affect everyone: sporotrichosis, or may affect only immunodeficient persons: mucormycosis. They do not spare Africa, we may encounter basidiobolomycosis, rhinophycomycosis, dermatophytosis, sporotrichosis and, more recently reported, rhinosporidiosis. Important therapeutic progresses have been accomplished with amphotericin B and with antifungus imidazole compounds (miconazole and ketoconazole). Surgical intervention is sometime recommended in chromomycosis and rhinosporidiosis.

  3. Deep penetration calculations

    International Nuclear Information System (INIS)

    Thompson, W.L.; Deutsch, O.L.; Booth, T.E.

    1980-04-01

    Several Monte Carlo techniques are compared in the transport of neutrons of different source energies through two different deep-penetration problems each with two parts. The first problem involves transmission through a 200-cm concrete slab. The second problem is a 90 0 bent pipe jacketed by concrete. In one case the pipe is void, and in the other it is filled with liquid sodium. Calculations are made with two different Los Alamos Monte Carlo codes: the continuous-energy code MCNP and the multigroup code MCMG

  4. Deep Super Learner: A Deep Ensemble for Classification Problems

    OpenAIRE

    Young, Steven; Abdou, Tamer; Bener, Ayse

    2018-01-01

    Deep learning has become very popular for tasks such as predictive modeling and pattern recognition in handling big data. Deep learning is a powerful machine learning method that extracts lower level features and feeds them forward for the next layer to identify higher level features that improve performance. However, deep neural networks have drawbacks, which include many hyper-parameters and infinite architectures, opaqueness into results, and relatively slower convergence on smaller datase...

  5. Deep sea biophysics

    International Nuclear Information System (INIS)

    Yayanos, A.A.

    1982-01-01

    A collection of deep-sea bacterial cultures was completed. Procedures were instituted to shelter the culture collection from accidential warming. A substantial data base on the rates of reproduction of more than 100 strains of bacteria from that collection was obtained from experiments and the analysis of that data was begun. The data on the rates of reproduction were obtained under conditions of temperature and pressure found in the deep sea. The experiments were facilitated by inexpensively fabricated pressure vessels, by the streamlining of the methods for the study of kinetics at high pressures, and by computer-assisted methods. A polybarothermostat was used to study the growth of bacteria along temperature gradients at eight distinct pressures. This device should allow for the study of microbial processes in the temperature field simulating the environment around buried HLW. It is small enough to allow placement in a radiation field in future studies. A flow fluorocytometer was fabricated. This device will be used to determine the DNA content per cell in bacteria grown in laboratory culture and in microorganisms in samples from the ocean. The technique will be tested for its rapidity in determining the concentration of cells (standing stock of microorganisms) in samples from the ocean

  6. Deep Learning in Radiology.

    Science.gov (United States)

    McBee, Morgan P; Awan, Omer A; Colucci, Andrew T; Ghobadi, Comeron W; Kadom, Nadja; Kansagra, Akash P; Tridandapani, Srini; Auffermann, William F

    2018-03-29

    As radiology is inherently a data-driven specialty, it is especially conducive to utilizing data processing techniques. One such technique, deep learning (DL), has become a remarkably powerful tool for image processing in recent years. In this work, the Association of University Radiologists Radiology Research Alliance Task Force on Deep Learning provides an overview of DL for the radiologist. This article aims to present an overview of DL in a manner that is understandable to radiologists; to examine past, present, and future applications; as well as to evaluate how radiologists may benefit from this remarkable new tool. We describe several areas within radiology in which DL techniques are having the most significant impact: lesion or disease detection, classification, quantification, and segmentation. The legal and ethical hurdles to implementation are also discussed. By taking advantage of this powerful tool, radiologists can become increasingly more accurate in their interpretations with fewer errors and spend more time to focus on patient care. Copyright © 2018 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.

  7. Deep Learning Microscopy

    KAUST Repository

    Rivenson, Yair

    2017-05-12

    We demonstrate that a deep neural network can significantly improve optical microscopy, enhancing its spatial resolution over a large field-of-view and depth-of-field. After its training, the only input to this network is an image acquired using a regular optical microscope, without any changes to its design. We blindly tested this deep learning approach using various tissue samples that are imaged with low-resolution and wide-field systems, where the network rapidly outputs an image with remarkably better resolution, matching the performance of higher numerical aperture lenses, also significantly surpassing their limited field-of-view and depth-of-field. These results are transformative for various fields that use microscopy tools, including e.g., life sciences, where optical microscopy is considered as one of the most widely used and deployed techniques. Beyond such applications, our presented approach is broadly applicable to other imaging modalities, also spanning different parts of the electromagnetic spectrum, and can be used to design computational imagers that get better and better as they continue to image specimen and establish new transformations among different modes of imaging.

  8. Deep Transfer Metric Learning.

    Science.gov (United States)

    Junlin Hu; Jiwen Lu; Yap-Peng Tan; Jie Zhou

    2016-12-01

    Conventional metric learning methods usually assume that the training and test samples are captured in similar scenarios so that their distributions are assumed to be the same. This assumption does not hold in many real visual recognition applications, especially when samples are captured across different data sets. In this paper, we propose a new deep transfer metric learning (DTML) method to learn a set of hierarchical nonlinear transformations for cross-domain visual recognition by transferring discriminative knowledge from the labeled source domain to the unlabeled target domain. Specifically, our DTML learns a deep metric network by maximizing the inter-class variations and minimizing the intra-class variations, and minimizing the distribution divergence between the source domain and the target domain at the top layer of the network. To better exploit the discriminative information from the source domain, we further develop a deeply supervised transfer metric learning (DSTML) method by including an additional objective on DTML, where the output of both the hidden layers and the top layer are optimized jointly. To preserve the local manifold of input data points in the metric space, we present two new methods, DTML with autoencoder regularization and DSTML with autoencoder regularization. Experimental results on face verification, person re-identification, and handwritten digit recognition validate the effectiveness of the proposed methods.

  9. Stable isotope geochemistry of deep sea cherts

    Energy Technology Data Exchange (ETDEWEB)

    Kolodny, Y; Epstein, S [California Inst. of Tech., Pasadena (USA). Div. of Geological Sciences

    1976-10-01

    Seventy four samples of DSDP (Deep Sea Drilling Project) recovered cherts of Jurassic to Miocene age from varying locations, and 27 samples of on-land exposed cherts were analyzed for the isotopic composition of their oxygen and hydrogen. These studies were accompanied by mineralogical analyses and some isotopic analyses of the coexisting carbonates. delta/sup 18/0 of chert ranges between 27 and 39 parts per thousand relative to SMOW, delta/sup 18/0 of porcellanite - between 30 and 42 parts per thousand. The consistent enrichment of opal-CT in porcellanites in /sup 18/0 with respect to coexisting microcrystalline quartz in chert is probably a reflection of a different temperature (depth) of diagenesis of the two phases. delta/sup 18/0 of deep sea cherts generally decrease with increasing age, indicating an overall cooling of the ocean bottom during the last 150 m.y. A comparison of this trend with that recorded by benthonic foraminifera (Douglas et al., Initial Reports of the Deep Sea Drilling Project; 32:509(1975)) indicates the possibility of delta/sup 18/0 in deep sea cherts not being frozen in until several tens of millions of years after deposition. Cherts of any Age show a spread of delta/sup 18/0 values, increasing diagenesis being reflected in a lowering of delta/sup 18/0. Drusy quartz has the lowest delta/sup 18/0 values. On land exposed cherts are consistently depleted in /sup 18/0 in comparison to their deep sea time equivalent cherts. Water extracted from deep sea cherts ranges between 0.5 and 1.4 wt%. deltaD of this water ranges between -78 and -95 parts per thousand and is not a function of delta/sup 18/0 of the cherts (or the temperature of their formation).

  10. Deep Reinforcement Learning: An Overview

    OpenAIRE

    Li, Yuxi

    2017-01-01

    We give an overview of recent exciting achievements of deep reinforcement learning (RL). We discuss six core elements, six important mechanisms, and twelve applications. We start with background of machine learning, deep learning and reinforcement learning. Next we discuss core RL elements, including value function, in particular, Deep Q-Network (DQN), policy, reward, model, planning, and exploration. After that, we discuss important mechanisms for RL, including attention and memory, unsuperv...

  11. Deep Feature Consistent Variational Autoencoder

    OpenAIRE

    Hou, Xianxu; Shen, Linlin; Sun, Ke; Qiu, Guoping

    2016-01-01

    We present a novel method for constructing Variational Autoencoder (VAE). Instead of using pixel-by-pixel loss, we enforce deep feature consistency between the input and the output of a VAE, which ensures the VAE's output to preserve the spatial correlation characteristics of the input, thus leading the output to have a more natural visual appearance and better perceptual quality. Based on recent deep learning works such as style transfer, we employ a pre-trained deep convolutional neural net...

  12. Deep learning for image classification

    Science.gov (United States)

    McCoppin, Ryan; Rizki, Mateen

    2014-06-01

    This paper provides an overview of deep learning and introduces the several subfields of deep learning including a specific tutorial of convolutional neural networks. Traditional methods for learning image features are compared to deep learning techniques. In addition, we present our preliminary classification results, our basic implementation of a convolutional restricted Boltzmann machine on the Mixed National Institute of Standards and Technology database (MNIST), and we explain how to use deep learning networks to assist in our development of a robust gender classification system.

  13. Deep learning? What deep learning? | Fourie | South African ...

    African Journals Online (AJOL)

    In teaching generally over the past twenty years, there has been a move towards teaching methods that encourage deep, rather than surface approaches to learning. The reason for this being that students, who adopt a deep approach to learning are considered to have learning outcomes of a better quality and desirability ...

  14. Deep sea radionuclides

    International Nuclear Information System (INIS)

    Kanisch, G.; Vobach, M.

    1993-01-01

    Every year since 1979, either in sping or in summer, the fishing research vessel 'Walther Herwig' goes to the North Atlantic disposal areas of solid radioactive wastes, and, for comparative purposes, to other areas, in order to collect water samples, plankton and nekton, and, from the deep sea bed, sediment samples and benthos organisms. In addition to data on the radionuclide contents of various media, information about the plankton, nekton and benthos organisms living in those areas and about their biomasses could be gathered. The investigations are aimed at acquiring scientifically founded knowledge of the uptake of radioactive substances by microorganisms, and their migration from the sea bottom to the areas used by man. (orig.) [de

  15. Deep inelastic phenomena

    International Nuclear Information System (INIS)

    Aubert, J.J.

    1982-01-01

    The experimental situation of the deep inelastic scattering for electrons (muons) is reviewed. A brief history of experimentation highlights Mohr and Nicoll's 1932 experiment on electron-atom scattering and Hofstadter's 1950 experiment on electron-nucleus scattering. The phenomenology of electron-nucleon scattering carried out between 1960 and 1970 is described, with emphasis on the parton model, and scaling. Experiments at SLAC and FNAL since 1974 exhibit scaling violations. Three muon-nucleon scattering experiments at BFP, BCDMA, and EMA, currently producing new results in the high Q 2 domain suggest a rather flat behaviour of the structure function at fixed x as a function of Q 2 . It is seen that the structure measured in DIS can then be projected into a pure hadronic process to predict a cross section. Protonneutron difference, moment analysis, and Drell-Yan pairs are also considered

  16. Perforating pilomatrixoma showing atypical presentation: A rare clinical variant

    Directory of Open Access Journals (Sweden)

    Nevra Seyhan

    2018-03-01

    Full Text Available Pilomatrixoma, also known as calcifying epithelioma of Malherbe, is a rare benign skin tumor arising from hair follicle stem cells. The most common localization is the head and neck region. Female/male ratio is 3/2. It shows deep subcutaneous placement and occurs in the first two decades of life. Its diameter ranges from 0.5 cm to 3 cm. Multiple lesions are rarely seen. Histopathologically it is characterized by basoloid and ghost cells. Perforating type is a rare clinical variant. Treatment is surgical excision. Our case is presented to draw attention to a rare clinical variant of pilomatrixioma.

  17. Inhibition of Blumeria graminis germination and germling development within colonies of oat mildew

    DEFF Research Database (Denmark)

    Carver, T.L.W.; Roberts, P.C.; Thomas, B.J.

    2001-01-01

    Germination by Blumeria graminis. DC Speer ff. spp. avenae, hordei and tritici, was greatly suppressed when conidia fell within colonies of ff. spp. avenae or hordei established on susceptible oat or barley, respectively. On healthy oat or barley, and when distant from powdery mildew, colonies. all...... ff. spp. formed normal appressoria. This was also true When conidia germinated within established barley mildew colonies. Within barley mildew colonies, appressoria of f. sp. hordei penetrated epidermal cells formed haustoria more frequently than appressoria distant from colonies. Similarly, ff. spp....... avenae and tritici, normally unable to infect barley. frequently penetrated epidermal cells subtending established barley mildew colonies. Thus, colony, establishment induced barley epidermal cell accessibility, even to non-pathogenic ff. spp, In contrast. when all three ff. spp. germinated within...

  18. Abnormal germling development by brown rust and powdery mildew on cer barley mutants

    NARCIS (Netherlands)

    Rubiales, D.; Ramirez, M.C.; Carver, T.L.W.; Niks, R.E.

    2001-01-01

    The barley leaf rust fungus forms appressoria over host leaf stomata and penetrates via the stomatal pore. High levels of avoidance to leaf rust fungi have been described in some wild accessions of Hordeum species where a prominent wax layer on the stomata inhibits triggering of fungal appressorium

  19. Deep-Sea Microbes: Linking Biogeochemical Rates to -Omics Approaches

    Science.gov (United States)

    Herndl, G. J.; Sintes, E.; Bayer, B.; Bergauer, K.; Amano, C.; Hansman, R.; Garcia, J.; Reinthaler, T.

    2016-02-01

    Over the past decade substantial progress has been made in determining deep ocean microbial activity and resolving some of the enigmas in understanding the deep ocean carbon flux. Also, metagenomics approaches have shed light onto the dark ocean's microbes but linking -omics approaches to biogeochemical rate measurements are generally rare in microbial oceanography and even more so for the deep ocean. In this presentation, we will show by combining metagenomics, -proteomics and biogeochemical rate measurements on the bulk and single-cell level that deep-sea microbes exhibit characteristics of generalists with a large genome repertoire, versatile in utilizing substrate as revealed by metaproteomics. This is in striking contrast with the apparently rather uniform dissolved organic matter pool in the deep ocean. Combining the different -omics approaches with metabolic rate measurements, we will highlight some major inconsistencies and enigmas in our understanding of the carbon cycling and microbial food web structure in the dark ocean.

  20. Multiscale deep features learning for land-use scene recognition

    Science.gov (United States)

    Yuan, Baohua; Li, Shijin; Li, Ning

    2018-01-01

    The features extracted from deep convolutional neural networks (CNNs) have shown their promise as generic descriptors for land-use scene recognition. However, most of the work directly adopts the deep features for the classification of remote sensing images, and does not encode the deep features for improving their discriminative power, which can affect the performance of deep feature representations. To address this issue, we propose an effective framework, LASC-CNN, obtained by locality-constrained affine subspace coding (LASC) pooling of a CNN filter bank. LASC-CNN obtains more discriminative deep features than directly extracted from CNNs. Furthermore, LASC-CNN builds on the top convolutional layers of CNNs, which can incorporate multiscale information and regions of arbitrary resolution and sizes. Our experiments have been conducted using two widely used remote sensing image databases, and the results show that the proposed method significantly improves the performance when compared to other state-of-the-art methods.

  1. Applying a punch with microridges in multistage deep drawing processes.

    Science.gov (United States)

    Lin, Bor-Tsuen; Yang, Cheng-Yu

    2016-01-01

    The developers of high aspect ratio components aim to minimize the processing stages in deep drawing processes. This study elucidates the application of microridge punches in multistage deep drawing processes. A microridge punch improves drawing performance, thereby reducing the number of stages required in deep forming processes. As an example, the original eight-stage deep forming process for a copper cylindrical cup with a high aspect ratio was analyzed by finite element simulation. Microridge punch designs were introduced in Stages 4 and 7 to replace the original punches. In addition, Stages 3 and 6 were eliminated. Finally, these changes were verified through experiments. The results showed that the microridge punches reduced the number of deep drawing stages yielding similar thickness difference percentages. Further, the numerical and experimental results demonstrated good consistency in the thickness distribution.

  2. AN EFFICIENT METHOD FOR DEEP WEB CRAWLER BASED ON ACCURACY -A REVIEW

    OpenAIRE

    Pranali Zade1, Dr.S.W.Mohod2

    2018-01-01

    As deep web grows at a very fast pace, there has been increased interest in techniques that help efficiently locate deep-web interfaces. However, due to the large volume of web resources and the dynamic nature of deep web, achieving wide coverage and high efficiency is a challenging issue. We propose a three-stage framework, for efficient harvesting deep web interfaces. Project experimental results on a set of representative domains show the agility and accuracy of our proposed crawler framew...

  3. Performance of deep geothermal energy systems

    Science.gov (United States)

    Manikonda, Nikhil

    Geothermal energy is an important source of clean and renewable energy. This project deals with the study of deep geothermal power plants for the generation of electricity. The design involves the extraction of heat from the Earth and its conversion into electricity. This is performed by allowing fluid deep into the Earth where it gets heated due to the surrounding rock. The fluid gets vaporized and returns to the surface in a heat pipe. Finally, the energy of the fluid is converted into electricity using turbine or organic rankine cycle (ORC). The main feature of the system is the employment of side channels to increase the amount of thermal energy extracted. A finite difference computer model is developed to solve the heat transport equation. The numerical model was employed to evaluate the performance of the design. The major goal was to optimize the output power as a function of parameters such as thermal diffusivity of the rock, depth of the main well, number and length of lateral channels. The sustainable lifetime of the system for a target output power of 2 MW has been calculated for deep geothermal systems with drilling depths of 8000 and 10000 meters, and a financial analysis has been performed to evaluate the economic feasibility of the system for a practical range of geothermal parameters. Results show promising an outlook for deep geothermal systems for practical applications.

  4. Deep learning for studies of galaxy morphology

    Science.gov (United States)

    Tuccillo, D.; Huertas-Company, M.; Decencière, E.; Velasco-Forero, S.

    2017-06-01

    Establishing accurate morphological measurements of galaxies in a reasonable amount of time for future big-data surveys such as EUCLID, the Large Synoptic Survey Telescope or the Wide Field Infrared Survey Telescope is a challenge. Because of its high level of abstraction with little human intervention, deep learning appears to be a promising approach. Deep learning is a rapidly growing discipline that models high-level patterns in data as complex multilayered networks. In this work we test the ability of deep convolutional networks to provide parametric properties of Hubble Space Telescope like galaxies (half-light radii, Sérsic indices, total flux etc..). We simulate a set of galaxies including point spread function and realistic noise from the CANDELS survey and try to recover the main galaxy parameters using deep-learning. We compare the results with the ones obtained with the commonly used profile fitting based software GALFIT. This way showing that with our method we obtain results at least equally good as the ones obtained with GALFIT but, once trained, with a factor 5 hundred time faster.

  5. Context and Deep Learning Design

    Science.gov (United States)

    Boyle, Tom; Ravenscroft, Andrew

    2012-01-01

    Conceptual clarification is essential if we are to establish a stable and deep discipline of technology enhanced learning. The technology is alluring; this can distract from deep design in a surface rush to exploit the affordances of the new technology. We need a basis for design, and a conceptual unit of organization, that are applicable across…

  6. A Survey: Time Travel in Deep Learning Space: An Introduction to Deep Learning Models and How Deep Learning Models Evolved from the Initial Ideas

    OpenAIRE

    Wang, Haohan; Raj, Bhiksha

    2015-01-01

    This report will show the history of deep learning evolves. It will trace back as far as the initial belief of connectionism modelling of brain, and come back to look at its early stage realization: neural networks. With the background of neural network, we will gradually introduce how convolutional neural network, as a representative of deep discriminative models, is developed from neural networks, together with many practical techniques that can help in optimization of neural networks. On t...

  7. Protein Secondary Structure Prediction Using Deep Convolutional Neural Fields.

    Science.gov (United States)

    Wang, Sheng; Peng, Jian; Ma, Jianzhu; Xu, Jinbo

    2016-01-11

    Protein secondary structure (SS) prediction is important for studying protein structure and function. When only the sequence (profile) information is used as input feature, currently the best predictors can obtain ~80% Q3 accuracy, which has not been improved in the past decade. Here we present DeepCNF (Deep Convolutional Neural Fields) for protein SS prediction. DeepCNF is a Deep Learning extension of Conditional Neural Fields (CNF), which is an integration of Conditional Random Fields (CRF) and shallow neural networks. DeepCNF can model not only complex sequence-structure relationship by a deep hierarchical architecture, but also interdependency between adjacent SS labels, so it is much more powerful than CNF. Experimental results show that DeepCNF can obtain ~84% Q3 accuracy, ~85% SOV score, and ~72% Q8 accuracy, respectively, on the CASP and CAMEO test proteins, greatly outperforming currently popular predictors. As a general framework, DeepCNF can be used to predict other protein structure properties such as contact number, disorder regions, and solvent accessibility.

  8. Learning with hierarchical-deep models.

    Science.gov (United States)

    Salakhutdinov, Ruslan; Tenenbaum, Joshua B; Torralba, Antonio

    2013-08-01

    We introduce HD (or “Hierarchical-Deep”) models, a new compositional learning architecture that integrates deep learning models with structured hierarchical Bayesian (HB) models. Specifically, we show how we can learn a hierarchical Dirichlet process (HDP) prior over the activities of the top-level features in a deep Boltzmann machine (DBM). This compound HDP-DBM model learns to learn novel concepts from very few training example by learning low-level generic features, high-level features that capture correlations among low-level features, and a category hierarchy for sharing priors over the high-level features that are typical of different kinds of concepts. We present efficient learning and inference algorithms for the HDP-DBM model and show that it is able to learn new concepts from very few examples on CIFAR-100 object recognition, handwritten character recognition, and human motion capture datasets.

  9. Application of AMT in detecting deep geological structures in Lejia district of Xiangshan uranium ore field

    International Nuclear Information System (INIS)

    Duan Shuxin; Liu Hu

    2014-01-01

    In recent years, exploration in Xiangshan uranium ore field shows that the intersection of faults and the interface of different rock formation and the basement is an important sign of deep ore- prospecting. In order to evaluate deep uranium resource in Lejia district, audio magnetotelluric method (AMT) was undertaken to carry out profile investigation. With that method, we discerned the interface of different rock formation and the basement successfully, and faults in the deep, which provides a good basis for the prediction of deep uranium resource. Drilling results show that AMT method has an obvious advantage in detecting deep geological structures in Xiangshan. (authors)

  10. Deep soft tissue leiomyoma of the thigh

    International Nuclear Information System (INIS)

    Watson, G.M.T.; Saifuddin, A.; Sandison, A.

    1999-01-01

    A case of ossified leiomyoma of the deep soft tissues of the left thigh is presented. The radiographic appearance suggested a low-grade chondrosarcoma. MRI of the lesion showed signal characteristics similar to muscle on both T1- and T2-weighted spin echo sequences with linear areas of high signal intensity on T1-weighted images consistent with medullary fat in metaplastic bone. Histopathological examination of the resected specimen revealed a benign ossified soft tissue leiomyoma. (orig.)

  11. Deep electroproduction of exotic hybrid mesons

    International Nuclear Information System (INIS)

    Anikin, I.V.; Pire, B.; Szymanowski, L.; Teryaev, O.V.; Wallon, S.

    2004-01-01

    We evaluate the leading order amplitude for the deep exclusive electroproduction of an exotic hybrid meson in the Bjorken regime. We show that, contrarily to naive expectation, this amplitude factorizes at the twist 2 level and thus scales like usual meson electroproduction when the virtual photon and the hybrid meson are longitudinally polarized. Exotic hybrid mesons may thus be studied in electroproduction experiments at JLAB, HERA (HERMES) or CERN (Compass)

  12. Deep Learning Fluid Mechanics

    Science.gov (United States)

    Barati Farimani, Amir; Gomes, Joseph; Pande, Vijay

    2017-11-01

    We have developed a new data-driven model paradigm for the rapid inference and solution of the constitutive equations of fluid mechanic by deep learning models. Using generative adversarial networks (GAN), we train models for the direct generation of solutions to steady state heat conduction and incompressible fluid flow without knowledge of the underlying governing equations. Rather than using artificial neural networks to approximate the solution of the constitutive equations, GANs can directly generate the solutions to these equations conditional upon an arbitrary set of boundary conditions. Both models predict temperature, velocity and pressure fields with great test accuracy (>99.5%). The application of our framework for inferring and generating the solutions of partial differential equations can be applied to any physical phenomena and can be used to learn directly from experiments where the underlying physical model is complex or unknown. We also have shown that our framework can be used to couple multiple physics simultaneously, making it amenable to tackle multi-physics problems.

  13. Deep space telescopes

    CERN Multimedia

    CERN. Geneva

    2006-01-01

    The short series of seminars will address results and aims of current and future space astrophysics as the cultural framework for the development of deep space telescopes. It will then present such new tools, as they are currently available to, or imagined by, the scientific community, in the context of the science plans of ESA and of all major world space agencies. Ground-based astronomy, in the 400 years since Galileo’s telescope, has given us a profound phenomenological comprehension of our Universe, but has traditionally been limited to the narrow band(s) to which our terrestrial atmosphere is transparent. Celestial objects, however, do not care about our limitations, and distribute most of the information about their physics throughout the complete electromagnetic spectrum. Such information is there for the taking, from millimiter wavelengths to gamma rays. Forty years astronomy from space, covering now most of the e.m. spectrum, have thus given us a better understanding of our physical Universe then t...

  14. Deep Mapping and Spatial Anthropology

    Directory of Open Access Journals (Sweden)

    Les Roberts

    2016-01-01

    Full Text Available This paper provides an introduction to the Humanities Special Issue on “Deep Mapping”. It sets out the rationale for the collection and explores the broad-ranging nature of perspectives and practices that fall within the “undisciplined” interdisciplinary domain of spatial humanities. Sketching a cross-current of ideas that have begun to coalesce around the concept of “deep mapping”, the paper argues that rather than attempting to outline a set of defining characteristics and “deep” cartographic features, a more instructive approach is to pay closer attention to the multivalent ways deep mapping is performatively put to work. Casting a critical and reflexive gaze over the developing discourse of deep mapping, it is argued that what deep mapping “is” cannot be reduced to the otherwise a-spatial and a-temporal fixity of the “deep map”. In this respect, as an undisciplined survey of this increasing expansive field of study and practice, the paper explores the ways in which deep mapping can engage broader discussion around questions of spatial anthropology.

  15. Deep bite malocclusion: exploration of the skeletal and dental factors

    International Nuclear Information System (INIS)

    Bhateja, N.K.; Fida, M.; Shaikh, A.

    2016-01-01

    Correction of deep bite is crucial for maintenance of dental hard and soft tissue structures and for prevention of temporomandibular joint disorders. Exploration of underlying skeletal and dental factors is essential for efficient and individualized treatment planning. To date etiological factors of dental and skeletal deep bite have not been explored in Pakistani orthodontic patients. The objectives of this study were to explore frequencies of dental and skeletal etiological factors in deep bite patients and to determine correlations amongst dental and skeletal etiological factors of deep bite. Methods: The study included a total of 113 subjects (males=35; females=78) with no craniofacial syndromes or prior orthodontic treatment. Pre-treatment orthodontic records were used to evaluate various dental and skeletal parameters. Descriptive statistics of each parameter were calculated. The various study parameters were correlated using Pearson's Correlation. Results: Deep curve of Spee was most frequently seen factor of dental deep bite (72.6%), followed by increased coronal length of upper incisors (28.3%), retroclined upper incisors (17.7%), retroclined lower incisors (8%) and increased coronal length of lower incisors (5.3%). Decreased gonial angle was most commonly found factor of skeletal deep bite (43.4%), followed by decreased mandibular plane angle (27.4%) and maxillary plane's clockwise rotation (26.5%). Frankfort mandibular plane angle and gonial angle showed a strong positive correlation (r=0.66, p=0.000). Conclusions: Reduced gonial angle is most frequently seen skeletal factor, signifying the importance of angulation and growth of ramus in development of deep bite. Deep curve of Spee is most frequently seen dental etiological component in deep bite subjects, hence signifying the importance of intruding the lower anterior teeth. (author)

  16. Deep learning for computational chemistry.

    Science.gov (United States)

    Goh, Garrett B; Hodas, Nathan O; Vishnu, Abhinav

    2017-06-15

    The rise and fall of artificial neural networks is well documented in the scientific literature of both computer science and computational chemistry. Yet almost two decades later, we are now seeing a resurgence of interest in deep learning, a machine learning algorithm based on multilayer neural networks. Within the last few years, we have seen the transformative impact of deep learning in many domains, particularly in speech recognition and computer vision, to the extent that the majority of expert practitioners in those field are now regularly eschewing prior established models in favor of deep learning models. In this review, we provide an introductory overview into the theory of deep neural networks and their unique properties that distinguish them from traditional machine learning algorithms used in cheminformatics. By providing an overview of the variety of emerging applications of deep neural networks, we highlight its ubiquity and broad applicability to a wide range of challenges in the field, including quantitative structure activity relationship, virtual screening, protein structure prediction, quantum chemistry, materials design, and property prediction. In reviewing the performance of deep neural networks, we observed a consistent outperformance against non-neural networks state-of-the-art models across disparate research topics, and deep neural network-based models often exceeded the "glass ceiling" expectations of their respective tasks. Coupled with the maturity of GPU-accelerated computing for training deep neural networks and the exponential growth of chemical data on which to train these networks on, we anticipate that deep learning algorithms will be a valuable tool for computational chemistry. © 2017 Wiley Periodicals, Inc. © 2017 Wiley Periodicals, Inc.

  17. Deep learning for computational chemistry

    Energy Technology Data Exchange (ETDEWEB)

    Goh, Garrett B. [Advanced Computing, Mathematics, and Data Division, Pacific Northwest National Laboratory, 902 Battelle Blvd Richland Washington 99354; Hodas, Nathan O. [Advanced Computing, Mathematics, and Data Division, Pacific Northwest National Laboratory, 902 Battelle Blvd Richland Washington 99354; Vishnu, Abhinav [Advanced Computing, Mathematics, and Data Division, Pacific Northwest National Laboratory, 902 Battelle Blvd Richland Washington 99354

    2017-03-08

    The rise and fall of artificial neural networks is well documented in the scientific literature of both the fields of computer science and computational chemistry. Yet almost two decades later, we are now seeing a resurgence of interest in deep learning, a machine learning algorithm based on “deep” neural networks. Within the last few years, we have seen the transformative impact of deep learning the computer science domain, notably in speech recognition and computer vision, to the extent that the majority of practitioners in those field are now regularly eschewing prior established models in favor of deep learning models. In this review, we provide an introductory overview into the theory of deep neural networks and their unique properties as compared to traditional machine learning algorithms used in cheminformatics. By providing an overview of the variety of emerging applications of deep neural networks, we highlight its ubiquity and broad applicability to a wide range of challenges in the field, including QSAR, virtual screening, protein structure modeling, QM calculations, materials synthesis and property prediction. In reviewing the performance of deep neural networks, we observed a consistent outperformance against non neural networks state-of-the-art models across disparate research topics, and deep neural network based models often exceeded the “glass ceiling” expectations of their respective tasks. Coupled with the maturity of GPU-accelerated computing for training deep neural networks and the exponential growth of chemical data on which to train these networks on, we anticipate that deep learning algorithms will be a useful tool and may grow into a pivotal role for various challenges in the computational chemistry field.

  18. DeepSimulator: a deep simulator for Nanopore sequencing

    KAUST Repository

    Li, Yu; Han, Renmin; Bi, Chongwei; Li, Mo; Wang, Sheng; Gao, Xin

    2017-01-01

    or assembled contigs, we simulate the electrical current signals by a context-dependent deep learning model, followed by a base-calling procedure to yield simulated reads. This workflow mimics the sequencing procedure more naturally. The thorough experiments

  19. DEWS (DEep White matter hyperintensity Segmentation framework): A fully automated pipeline for detecting small deep white matter hyperintensities in migraineurs.

    Science.gov (United States)

    Park, Bo-Yong; Lee, Mi Ji; Lee, Seung-Hak; Cha, Jihoon; Chung, Chin-Sang; Kim, Sung Tae; Park, Hyunjin

    2018-01-01

    Migraineurs show an increased load of white matter hyperintensities (WMHs) and more rapid deep WMH progression. Previous methods for WMH segmentation have limited efficacy to detect small deep WMHs. We developed a new fully automated detection pipeline, DEWS (DEep White matter hyperintensity Segmentation framework), for small and superficially-located deep WMHs. A total of 148 non-elderly subjects with migraine were included in this study. The pipeline consists of three components: 1) white matter (WM) extraction, 2) WMH detection, and 3) false positive reduction. In WM extraction, we adjusted the WM mask to re-assign misclassified WMHs back to WM using many sequential low-level image processing steps. In WMH detection, the potential WMH clusters were detected using an intensity based threshold and region growing approach. For false positive reduction, the detected WMH clusters were classified into final WMHs and non-WMHs using the random forest (RF) classifier. Size, texture, and multi-scale deep features were used to train the RF classifier. DEWS successfully detected small deep WMHs with a high positive predictive value (PPV) of 0.98 and true positive rate (TPR) of 0.70 in the training and test sets. Similar performance of PPV (0.96) and TPR (0.68) was attained in the validation set. DEWS showed a superior performance in comparison with other methods. Our proposed pipeline is freely available online to help the research community in quantifying deep WMHs in non-elderly adults.

  20. Deep UV LEDs

    Science.gov (United States)

    Han, Jung; Amano, Hiroshi; Schowalter, Leo

    2014-06-01

    Deep ultraviolet (DUV) photons interact strongly with a broad range of chemical and biological molecules; compact DUV light sources could enable a wide range of applications in chemi/bio-sensing, sterilization, agriculture, and industrial curing. The much shorter wavelength also results in useful characteristics related to optical diffraction (for lithography) and scattering (non-line-of-sight communication). The family of III-N (AlGaInN) compound semiconductors offers a tunable energy gap from infrared to DUV. While InGaN-based blue light emitters have been the primary focus for the obvious application of solid state lighting, there is a growing interest in the development of efficient UV and DUV light-emitting devices. In the past few years we have witnessed an increasing investment from both government and industry sectors to further the state of DUV light-emitting devices. The contributions in Semiconductor Science and Technology 's special issue on DUV devices provide an up-to-date snapshot covering many relevant topics in this field. Given the expected importance of bulk AlN substrate in DUV technology, we are pleased to include a review article by Hartmann et al on the growth of AlN bulk crystal by physical vapour transport. The issue of polarization field within the deep ultraviolet LEDs is examined in the article by Braut et al. Several commercial companies provide useful updates in their development of DUV emitters, including Nichia (Fujioka et al ), Nitride Semiconductors (Muramoto et al ) and Sensor Electronic Technology (Shatalov et al ). We believe these articles will provide an excellent overview of the state of technology. The growth of AlGaN heterostructures by molecular beam epitaxy, in contrast to the common organo-metallic vapour phase epitaxy, is discussed by Ivanov et al. Since hexagonal boron nitride (BN) has received much attention as both a UV and a two-dimensional electronic material, we believe it serves readers well to include the

  1. DEEP INFILTRATING ENDOMETRIOSIS

    Directory of Open Access Journals (Sweden)

    Martina Ribič-Pucelj

    2018-02-01

    Full Text Available Background: Endometriosis is not considered a unified disease, but a disease encompassing three differ- ent forms differentiated by aetiology and pathogenesis: peritoneal endometriosis, ovarian endometriosis and deep infiltrating endometriosis (DIE. The disease is classified as DIE when the lesions penetrate 5 mm or more into the retroperitoneal space. The estimated incidence of endometriosis in women of reproductive age ranges from 10–15 % and that of DIE from 3–10 %, the highest being in infertile women and in those with chronic pelvic pain. The leading symptoms of DIE are chronic pelvic pain which increases with age and correlates with the depth of infiltration and infertility. The most important diagnostic procedures are patient’s history and proper gynecological examination. The diagnosis is confirmed with laparoscopy. DIE can affect, beside reproductive organs, also bowel, bladder and ureters, therefore adi- tional diagnostic procedures must be performed preopertively to confirm or to exclude the involvement of the mentioned organs. Endometriosis is hormon dependent disease, there- fore several hormonal treatment regims are used to supress estrogen production but the symptoms recurr soon after caesation of the treatment. At the moment, surgical treatment with excision of all lesions, including those of bowel, bladder and ureters, is the method of choice but requires frequently interdisciplinary approach. Surgical treatment significantly reduces pain and improves fertility in inferile patients. Conclusions: DIE is not a rare form of endometriosis characterized by chronic pelvic pain and infertility. Medical treatment is not efficient. The method of choice is surgical treatment with excision of all lesions. It significantly reduces pelvic pain and enables high spontaneus and IVF preg- nacy rates.Therefore such patients should be treated at centres with experience in treatment of DIE and with possibility of interdisciplinary approach.

  2. Optimization of lining design in deep clays

    International Nuclear Information System (INIS)

    Rousset, G.; Bublitz, D.

    1989-01-01

    The main features of the mechanical behaviour of deep clay are time dependent effects and also the existence of a long term cohesion which may be taken into account for dimensioning galleries. In this text, a lining optimization test is presented. It concerns a gallery driven in deep clay, 230 m. deep, at Mol (Belgium). We show that sliding rib lining gives both: - an optimal tunnel face advance speed, a minimal closure of the gallery wall before setting the lining and therefore less likelihood of failure developing inside the rock mass. - limitation of the length of the non-lined part of the gallery. The chosen process allows on one hand the preservation of the rock mass integrity, and, on the other, use of the confinement effect to allow closure under high average stress conditions; this process can be considered as an optimal application of the convergence-confinement method. An important set of measurement devices is then presented along with results obtained for one year's operation. We show in particular that stress distribution in the lining is homogeneous and that the sliding limit can be measured with high precision

  3. Ensemble Network Architecture for Deep Reinforcement Learning

    Directory of Open Access Journals (Sweden)

    Xi-liang Chen

    2018-01-01

    Full Text Available The popular deep Q learning algorithm is known to be instability because of the Q-value’s shake and overestimation action values under certain conditions. These issues tend to adversely affect their performance. In this paper, we develop the ensemble network architecture for deep reinforcement learning which is based on value function approximation. The temporal ensemble stabilizes the training process by reducing the variance of target approximation error and the ensemble of target values reduces the overestimate and makes better performance by estimating more accurate Q-value. Our results show that this architecture leads to statistically significant better value evaluation and more stable and better performance on several classical control tasks at OpenAI Gym environment.

  4. Gas Classification Using Deep Convolutional Neural Networks

    Science.gov (United States)

    Peng, Pai; Zhao, Xiaojin; Pan, Xiaofang; Ye, Wenbin

    2018-01-01

    In this work, we propose a novel Deep Convolutional Neural Network (DCNN) tailored for gas classification. Inspired by the great success of DCNN in the field of computer vision, we designed a DCNN with up to 38 layers. In general, the proposed gas neural network, named GasNet, consists of: six convolutional blocks, each block consist of six layers; a pooling layer; and a fully-connected layer. Together, these various layers make up a powerful deep model for gas classification. Experimental results show that the proposed DCNN method is an effective technique for classifying electronic nose data. We also demonstrate that the DCNN method can provide higher classification accuracy than comparable Support Vector Machine (SVM) methods and Multiple Layer Perceptron (MLP). PMID:29316723

  5. Gas Classification Using Deep Convolutional Neural Networks.

    Science.gov (United States)

    Peng, Pai; Zhao, Xiaojin; Pan, Xiaofang; Ye, Wenbin

    2018-01-08

    In this work, we propose a novel Deep Convolutional Neural Network (DCNN) tailored for gas classification. Inspired by the great success of DCNN in the field of computer vision, we designed a DCNN with up to 38 layers. In general, the proposed gas neural network, named GasNet, consists of: six convolutional blocks, each block consist of six layers; a pooling layer; and a fully-connected layer. Together, these various layers make up a powerful deep model for gas classification. Experimental results show that the proposed DCNN method is an effective technique for classifying electronic nose data. We also demonstrate that the DCNN method can provide higher classification accuracy than comparable Support Vector Machine (SVM) methods and Multiple Layer Perceptron (MLP).

  6. DAPs: Deep Action Proposals for Action Understanding

    KAUST Repository

    Escorcia, Victor

    2016-09-17

    Object proposals have contributed significantly to recent advances in object understanding in images. Inspired by the success of this approach, we introduce Deep Action Proposals (DAPs), an effective and efficient algorithm for generating temporal action proposals from long videos. We show how to take advantage of the vast capacity of deep learning models and memory cells to retrieve from untrimmed videos temporal segments, which are likely to contain actions. A comprehensive evaluation indicates that our approach outperforms previous work on a large scale action benchmark, runs at 134 FPS making it practical for large-scale scenarios, and exhibits an appealing ability to generalize, i.e. to retrieve good quality temporal proposals of actions unseen in training.

  7. Telepresence for Deep Space Missions

    Data.gov (United States)

    National Aeronautics and Space Administration — Incorporating telepresence technologies into deep space mission operations can give the crew and ground personnel the impression that they are in a location at time...

  8. Hybrid mask for deep etching

    KAUST Repository

    Ghoneim, Mohamed T.

    2017-01-01

    Deep reactive ion etching is essential for creating high aspect ratio micro-structures for microelectromechanical systems, sensors and actuators, and emerging flexible electronics. A novel hybrid dual soft/hard mask bilayer may be deposited during

  9. pDeep: Predicting MS/MS Spectra of Peptides with Deep Learning.

    Science.gov (United States)

    Zhou, Xie-Xuan; Zeng, Wen-Feng; Chi, Hao; Luo, Chunjie; Liu, Chao; Zhan, Jianfeng; He, Si-Min; Zhang, Zhifei

    2017-12-05

    In tandem mass spectrometry (MS/MS)-based proteomics, search engines rely on comparison between an experimental MS/MS spectrum and the theoretical spectra of the candidate peptides. Hence, accurate prediction of the theoretical spectra of peptides appears to be particularly important. Here, we present pDeep, a deep neural network-based model for the spectrum prediction of peptides. Using the bidirectional long short-term memory (BiLSTM), pDeep can predict higher-energy collisional dissociation, electron-transfer dissociation, and electron-transfer and higher-energy collision dissociation MS/MS spectra of peptides with >0.9 median Pearson correlation coefficients. Further, we showed that intermediate layer of the neural network could reveal physicochemical properties of amino acids, for example the similarities of fragmentation behaviors between amino acids. We also showed the potential of pDeep to distinguish extremely similar peptides (peptides that contain isobaric amino acids, for example, GG = N, AG = Q, or even I = L), which were very difficult to distinguish using traditional search engines.

  10. Deep Learning and Bayesian Methods

    OpenAIRE

    Prosper Harrison B.

    2017-01-01

    A revolution is underway in which deep neural networks are routinely used to solve diffcult problems such as face recognition and natural language understanding. Particle physicists have taken notice and have started to deploy these methods, achieving results that suggest a potentially significant shift in how data might be analyzed in the not too distant future. We discuss a few recent developments in the application of deep neural networks and then indulge in speculation about how such meth...

  11. Density functionals from deep learning

    OpenAIRE

    McMahon, Jeffrey M.

    2016-01-01

    Density-functional theory is a formally exact description of a many-body quantum system in terms of its density; in practice, however, approximations to the universal density functional are required. In this work, a model based on deep learning is developed to approximate this functional. Deep learning allows computational models that are capable of naturally discovering intricate structure in large and/or high-dimensional data sets, with multiple levels of abstraction. As no assumptions are ...

  12. Deep Charging Evaluation of Satellite Power and Communication System Components

    Science.gov (United States)

    Schneider, T. A.; Vaughn, J. A.; Chu, B.; Wong, F.; Gardiner, G.; Wright, K. H.; Phillips, B.

    2016-01-01

    Deep charging, in contrast to surface charging, focuses on electron penetration deep into insulating materials applied over conductors. A classic example of this scenario is an insulated wire. Deep charging can pose a threat to material integrity, and to sensitive electronics, when it gives rise to an electrostatic discharge or arc. With the advent of Electric Orbit Raising, which requires spiraling through Earth's radiation belts, satellites are subjected to high energy electron environments which they normally would not encounter. Beyond Earth orbit, missions to Jupiter and Saturn face deep charging concerns due to the high energy radiation environments. While predictions can be made about charging in insulating materials, it is difficult to extend those predictions to complicated geometries, such as the case of an insulating coating around a small wire, or a non-uniform silicone grouting on a bus bar. Therefore, to conclusively determine the susceptibility of a system to arcs from deep charging, experimental investigations must be carried out. This paper will describe the evaluation carried out by NASA's Marshall Space Flight Center on subscale flight-like samples developed by Space Systems/Loral, LLC. Specifically, deep charging evaluations of solar array wire coupons, a photovoltaic cell coupon, and a coaxial microwave transmission cable, will be discussed. The results of each evaluation will be benchmarked against control sample tests, as well as typical power system levels, to show no significant deep charging threat existed for this set of samples under the conditions tested.

  13. Deep surface rolling for fatigue life enhancement of laser clad aircraft aluminium alloy

    Energy Technology Data Exchange (ETDEWEB)

    Zhuang, W., E-mail: wyman.zhuang@dsto.defence.gov.au [Aerospace Division, Defence Science and Technology Organisation, 506 Lorimer Street, Fishermans Bend, Victoria 3207 (Australia); Liu, Q.; Djugum, R.; Sharp, P.K. [Aerospace Division, Defence Science and Technology Organisation, 506 Lorimer Street, Fishermans Bend, Victoria 3207 (Australia); Paradowska, A. [Australian Nuclear Science and Technology Organisation, Lucas Heights, NSW 2232 (Australia)

    2014-11-30

    Highlights: • Deep surface rolling as a post-repair enhancement technology was applied to the laser cladded 7075-T651 aluminium alloy specimens that simulated corrosion damage blend-out repair. • The residual stresses induced by the deep surface rolling process were measured. • The deep surface rolling process can introduce deep and high magnitude compressive residual stresses beyond the laser clad and substrate interface. • Spectrum fatigue test showed the fatigue life was significantly increased by deep surface rolling. - Abstract: Deep surface rolling can introduce deep compressive residual stresses into the surface of aircraft metallic structure to extend its fatigue life. To develop cost-effective aircraft structural repair technologies such as laser cladding, deep surface rolling was considered as an advanced post-repair surface enhancement technology. In this study, aluminium alloy 7075-T651 specimens with a blend-out region were first repaired using laser cladding technology. The surface of the laser cladding region was then treated by deep surface rolling. Fatigue testing was subsequently conducted for the laser clad, deep surface rolled and post-heat treated laser clad specimens. It was found that deep surface rolling can significantly improve the fatigue life in comparison with the laser clad baseline repair. In addition, three dimensional residual stresses were measured using neutron diffraction techniques. The results demonstrate that beneficial compressive residual stresses induced by deep surface rolling can reach considerable depths (more than 1.0 mm) below the laser clad surface.

  14. Deep surface rolling for fatigue life enhancement of laser clad aircraft aluminium alloy

    International Nuclear Information System (INIS)

    Zhuang, W.; Liu, Q.; Djugum, R.; Sharp, P.K.; Paradowska, A.

    2014-01-01

    Highlights: • Deep surface rolling as a post-repair enhancement technology was applied to the laser cladded 7075-T651 aluminium alloy specimens that simulated corrosion damage blend-out repair. • The residual stresses induced by the deep surface rolling process were measured. • The deep surface rolling process can introduce deep and high magnitude compressive residual stresses beyond the laser clad and substrate interface. • Spectrum fatigue test showed the fatigue life was significantly increased by deep surface rolling. - Abstract: Deep surface rolling can introduce deep compressive residual stresses into the surface of aircraft metallic structure to extend its fatigue life. To develop cost-effective aircraft structural repair technologies such as laser cladding, deep surface rolling was considered as an advanced post-repair surface enhancement technology. In this study, aluminium alloy 7075-T651 specimens with a blend-out region were first repaired using laser cladding technology. The surface of the laser cladding region was then treated by deep surface rolling. Fatigue testing was subsequently conducted for the laser clad, deep surface rolled and post-heat treated laser clad specimens. It was found that deep surface rolling can significantly improve the fatigue life in comparison with the laser clad baseline repair. In addition, three dimensional residual stresses were measured using neutron diffraction techniques. The results demonstrate that beneficial compressive residual stresses induced by deep surface rolling can reach considerable depths (more than 1.0 mm) below the laser clad surface

  15. How Stressful Is "Deep Bubbling"?

    Science.gov (United States)

    Tyrmi, Jaana; Laukkanen, Anne-Maria

    2017-03-01

    Water resistance therapy by phonating through a tube into the water is used to treat dysphonia. Deep submersion (≥10 cm in water, "deep bubbling") is used for hypofunctional voice disorders. Using it with caution is recommended to avoid vocal overloading. This experimental study aimed to investigate how strenuous "deep bubbling" is. Fourteen subjects, half of them with voice training, repeated the syllable [pa:] in comfortable speaking pitch and loudness, loudly, and in strained voice. Thereafter, they phonated a vowel-like sound both in comfortable loudness and loudly into a glass resonance tube immersed 10 cm into the water. Oral pressure, contact quotient (CQ, calculated from electroglottographic signal), and sound pressure level were studied. The peak oral pressure P(oral) during [p] and shuttering of the outer end of the tube was measured to estimate the subglottic pressure P(sub) and the mean P(oral) during vowel portions to enable calculation of transglottic pressure P(trans). Sensations during phonation were reported with an open-ended interview. P(sub) and P(oral) were higher in "deep bubbling" and P(trans) lower than in loud syllable phonation, but the CQ did not differ significantly. Similar results were obtained for the comparison between loud "deep bubbling" and strained phonation, although P(sub) did not differ significantly. Most of the subjects reported "deep bubbling" to be stressful only for respiratory and lip muscles. No big differences were found between trained and untrained subjects. The CQ values suggest that "deep bubbling" may increase vocal fold loading. Further studies should address impact stress during water resistance exercises. Copyright © 2017 The Voice Foundation. Published by Elsevier Inc. All rights reserved.

  16. Quantitative phenotyping via deep barcode sequencing.

    Science.gov (United States)

    Smith, Andrew M; Heisler, Lawrence E; Mellor, Joseph; Kaper, Fiona; Thompson, Michael J; Chee, Mark; Roth, Frederick P; Giaever, Guri; Nislow, Corey

    2009-10-01

    Next-generation DNA sequencing technologies have revolutionized diverse genomics applications, including de novo genome sequencing, SNP detection, chromatin immunoprecipitation, and transcriptome analysis. Here we apply deep sequencing to genome-scale fitness profiling to evaluate yeast strain collections in parallel. This method, Barcode analysis by Sequencing, or "Bar-seq," outperforms the current benchmark barcode microarray assay in terms of both dynamic range and throughput. When applied to a complex chemogenomic assay, Bar-seq quantitatively identifies drug targets, with performance superior to the benchmark microarray assay. We also show that Bar-seq is well-suited for a multiplex format. We completely re-sequenced and re-annotated the yeast deletion collection using deep sequencing, found that approximately 20% of the barcodes and common priming sequences varied from expectation, and used this revised list of barcode sequences to improve data quality. Together, this new assay and analysis routine provide a deep-sequencing-based toolkit for identifying gene-environment interactions on a genome-wide scale.

  17. Stellar Atmospheric Parameterization Based on Deep Learning

    Science.gov (United States)

    Pan, Ru-yang; Li, Xiang-ru

    2017-07-01

    Deep learning is a typical learning method widely studied in the fields of machine learning, pattern recognition, and artificial intelligence. This work investigates the problem of stellar atmospheric parameterization by constructing a deep neural network with five layers, and the node number in each layer of the network is respectively 3821-500-100-50-1. The proposed scheme is verified on both the real spectra measured by the Sloan Digital Sky Survey (SDSS) and the theoretic spectra computed with the Kurucz's New Opacity Distribution Function (NEWODF) model, to make an automatic estimation for three physical parameters: the effective temperature (Teff), surface gravitational acceleration (lg g), and metallic abundance (Fe/H). The results show that the stacked autoencoder deep neural network has a better accuracy for the estimation. On the SDSS spectra, the mean absolute errors (MAEs) are 79.95 for Teff/K, 0.0058 for (lg Teff/K), 0.1706 for lg (g/(cm·s-2)), and 0.1294 dex for the [Fe/H], respectively; On the theoretic spectra, the MAEs are 15.34 for Teff/K, 0.0011 for lg (Teff/K), 0.0214 for lg(g/(cm · s-2)), and 0.0121 dex for [Fe/H], respectively.

  18. Deep-inelastic electron-proton diffraction

    International Nuclear Information System (INIS)

    Dainton, J.B.

    1995-11-01

    Recent measurements by the H1 collaboration at HERA of the cross section for deep-inelastic electron-proton scattering in which the proton interacts with minimal energy transfer and limited 4-momentum transfer squared are presented in the form of the contribution F 2 D(3) to the proton structure function F 2 . By parametrising the cross section phenomenologically in terms of a leading effective Regge pole exchange and comparing the result with a similar parametrisation of hadronic pp physics, the proton interaction is demonstrated to be dominantly of a diffractive nature. The quantitative interpretation of the parametrisation in terms of the properties of an effective leading Regge pole exchange, the pomeron (IP), shows that there is no evidence for a 'harder' BFKL-motivated IP in such deep-inelastic proton diffraction. The total contribution of proton diffraction to deep-inelastic electron-proton scattering is measured to be ∝10% and to be rather insensitive to Bjorken-x and Q 2 . A first measurement of the partonic structure of diffractive exchange is presented. It is shown to be readily interpreted in terms of the exchange of gluons, and to suggest that the bulk of diffractive momentum transfer is carried by a leading gluon. (orig.)

  19. Bidirectional Nonnegative Deep Model and Its Optimization in Learning

    Directory of Open Access Journals (Sweden)

    Xianhua Zeng

    2016-01-01

    Full Text Available Nonnegative matrix factorization (NMF has been successfully applied in signal processing as a simple two-layer nonnegative neural network. Projective NMF (PNMF with fewer parameters was proposed, which projects a high-dimensional nonnegative data onto a lower-dimensional nonnegative subspace. Although PNMF overcomes the problem of out-of-sample of NMF, it does not consider the nonlinear characteristic of data and is only a kind of narrow signal decomposition method. In this paper, we combine the PNMF with deep learning and nonlinear fitting to propose a bidirectional nonnegative deep learning (BNDL model and its optimization learning algorithm, which can obtain nonlinear multilayer deep nonnegative feature representation. Experiments show that the proposed model can not only solve the problem of out-of-sample of NMF but also learn hierarchical nonnegative feature representations with better clustering performance than classical NMF, PNMF, and Deep Semi-NMF algorithms.

  20. Deep learning—Accelerating Next Generation Performance Analysis Systems?

    Directory of Open Access Journals (Sweden)

    Heike Brock

    2018-02-01

    Full Text Available Deep neural network architectures show superior performance in recognition and prediction tasks of the image, speech and natural language domains. The success of such multi-layered networks encourages their implementation in further application scenarios as the retrieval of relevant motion information for performance enhancement in sports. However, to date deep learning is only seldom applied to activity recognition problems of the human motion domain. Therefore, its use for sports data analysis might remain abstract to many practitioners. This paper provides a survey on recent works in the field of high-performance motion data and examines relevant technologies for subsequent deployment in real training systems. In particular, it discusses aspects of data acquisition, processing and network modeling. Analysis suggests the advantage of deep neural networks under difficult and noisy data conditions. However, further research is necessary to confirm the benefit of deep learning for next generation performance analysis systems.

  1. Dilution limits dissolved organic carbon utilization in the deep ocean

    KAUST Repository

    Arrieta, Jesus

    2015-03-19

    Oceanic dissolved organic carbon (DOC) is the second largest reservoir of organic carbon in the biosphere. About 72% of the global DOC inventory is stored in deep oceanic layers for years to centuries, supporting the current view that it consists of materials resistant to microbial degradation. An alternative hypothesis is that deep-water DOC consists of many different, intrinsically labile compounds at concentrations too low to compensate for the metabolic costs associated to their utilization. Here, we present experimental evidence showing that low concentrations rather than recalcitrance preclude consumption of a substantial fraction of DOC, leading to slow microbial growth in the deep ocean. These findings demonstrate an alternative mechanism for the long-term storage of labile DOC in the deep ocean, which has been hitherto largely ignored. © 2015, American Association for the Advancement of Science. All rights reserved.

  2. Dilution limits dissolved organic carbon utilization in the deep ocean

    KAUST Repository

    Arrieta, J M; Mayol, Eva; Hansman, Roberta L.; Herndl, Gerhard J.; Dittmar, Thorsten; Duarte, Carlos M.

    2015-01-01

    Oceanic dissolved organic carbon (DOC) is the second largest reservoir of organic carbon in the biosphere. About 72% of the global DOC inventory is stored in deep oceanic layers for years to centuries, supporting the current view that it consists of materials resistant to microbial degradation. An alternative hypothesis is that deep-water DOC consists of many different, intrinsically labile compounds at concentrations too low to compensate for the metabolic costs associated to their utilization. Here, we present experimental evidence showing that low concentrations rather than recalcitrance preclude consumption of a substantial fraction of DOC, leading to slow microbial growth in the deep ocean. These findings demonstrate an alternative mechanism for the long-term storage of labile DOC in the deep ocean, which has been hitherto largely ignored. © 2015, American Association for the Advancement of Science. All rights reserved.

  3. Blind source deconvolution for deep Earth seismology

    Science.gov (United States)

    Stefan, W.; Renaut, R.; Garnero, E. J.; Lay, T.

    2007-12-01

    We present an approach to automatically estimate an empirical source characterization of deep earthquakes recorded teleseismically and subsequently remove the source from the recordings by applying regularized deconvolution. A principle goal in this work is to effectively deblur the seismograms, resulting in more impulsive and narrower pulses, permitting better constraints in high resolution waveform analyses. Our method consists of two stages: (1) we first estimate the empirical source by automatically registering traces to their 1st principal component with a weighting scheme based on their deviation from this shape, we then use this shape as an estimation of the earthquake source. (2) We compare different deconvolution techniques to remove the source characteristic from the trace. In particular Total Variation (TV) regularized deconvolution is used which utilizes the fact that most natural signals have an underlying spareness in an appropriate basis, in this case, impulsive onsets of seismic arrivals. We show several examples of deep focus Fiji-Tonga region earthquakes for the phases S and ScS, comparing source responses for the separate phases. TV deconvolution is compared to the water level deconvolution, Tikenov deconvolution, and L1 norm deconvolution, for both data and synthetics. This approach significantly improves our ability to study subtle waveform features that are commonly masked by either noise or the earthquake source. Eliminating source complexities improves our ability to resolve deep mantle triplications, waveform complexities associated with possible double crossings of the post-perovskite phase transition, as well as increasing stability in waveform analyses used for deep mantle anisotropy measurements.

  4. Gene expression inference with deep learning.

    Science.gov (United States)

    Chen, Yifei; Li, Yi; Narayan, Rajiv; Subramanian, Aravind; Xie, Xiaohui

    2016-06-15

    Large-scale gene expression profiling has been widely used to characterize cellular states in response to various disease conditions, genetic perturbations, etc. Although the cost of whole-genome expression profiles has been dropping steadily, generating a compendium of expression profiling over thousands of samples is still very expensive. Recognizing that gene expressions are often highly correlated, researchers from the NIH LINCS program have developed a cost-effective strategy of profiling only ∼1000 carefully selected landmark genes and relying on computational methods to infer the expression of remaining target genes. However, the computational approach adopted by the LINCS program is currently based on linear regression (LR), limiting its accuracy since it does not capture complex nonlinear relationship between expressions of genes. We present a deep learning method (abbreviated as D-GEX) to infer the expression of target genes from the expression of landmark genes. We used the microarray-based Gene Expression Omnibus dataset, consisting of 111K expression profiles, to train our model and compare its performance to those from other methods. In terms of mean absolute error averaged across all genes, deep learning significantly outperforms LR with 15.33% relative improvement. A gene-wise comparative analysis shows that deep learning achieves lower error than LR in 99.97% of the target genes. We also tested the performance of our learned model on an independent RNA-Seq-based GTEx dataset, which consists of 2921 expression profiles. Deep learning still outperforms LR with 6.57% relative improvement, and achieves lower error in 81.31% of the target genes. D-GEX is available at https://github.com/uci-cbcl/D-GEX CONTACT: xhx@ics.uci.edu Supplementary data are available at Bioinformatics online. © The Author 2016. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com.

  5. DeepVel: Deep learning for the estimation of horizontal velocities at the solar surface

    Science.gov (United States)

    Asensio Ramos, A.; Requerey, I. S.; Vitas, N.

    2017-07-01

    Many phenomena taking place in the solar photosphere are controlled by plasma motions. Although the line-of-sight component of the velocity can be estimated using the Doppler effect, we do not have direct spectroscopic access to the components that are perpendicular to the line of sight. These components are typically estimated using methods based on local correlation tracking. We have designed DeepVel, an end-to-end deep neural network that produces an estimation of the velocity at every single pixel, every time step, and at three different heights in the atmosphere from just two consecutive continuum images. We confront DeepVel with local correlation tracking, pointing out that they give very similar results in the time and spatially averaged cases. We use the network to study the evolution in height of the horizontal velocity field in fragmenting granules, supporting the buoyancy-braking mechanism for the formation of integranular lanes in these granules. We also show that DeepVel can capture very small vortices, so that we can potentially expand the scaling cascade of vortices to very small sizes and durations. The movie attached to Fig. 3 is available at http://www.aanda.org

  6. Accelerating Deep Learning with Shrinkage and Recall

    OpenAIRE

    Zheng, Shuai; Vishnu, Abhinav; Ding, Chris

    2016-01-01

    Deep Learning is a very powerful machine learning model. Deep Learning trains a large number of parameters for multiple layers and is very slow when data is in large scale and the architecture size is large. Inspired from the shrinking technique used in accelerating computation of Support Vector Machines (SVM) algorithm and screening technique used in LASSO, we propose a shrinking Deep Learning with recall (sDLr) approach to speed up deep learning computation. We experiment shrinking Deep Lea...

  7. What Really is Deep Learning Doing?

    OpenAIRE

    Xiong, Chuyu

    2017-01-01

    Deep learning has achieved a great success in many areas, from computer vision to natural language processing, to game playing, and much more. Yet, what deep learning is really doing is still an open question. There are a lot of works in this direction. For example, [5] tried to explain deep learning by group renormalization, and [6] tried to explain deep learning from the view of functional approximation. In order to address this very crucial question, here we see deep learning from perspect...

  8. Deep Water Ocean Acoustics

    Science.gov (United States)

    2016-12-22

    roughly 28°S. The second is the Hawaiian Island Chain, extending to Midway Island at 28°N, 177°W and finally the Emperor Seamount chain running due...dimension array centered near Ascension. The climatology ocean (WOA09) showed very little seasonal dependence or change from the geodesic and this is

  9. pathways to deep decarbonization - 2014 report

    International Nuclear Information System (INIS)

    Sachs, Jeffrey; Guerin, Emmanuel; Mas, Carl; Schmidt-Traub, Guido; Tubiana, Laurence; Waisman, Henri; Colombier, Michel; Bulger, Claire; Sulakshana, Elana; Zhang, Kathy; Barthelemy, Pierre; Spinazze, Lena; Pharabod, Ivan

    2014-09-01

    The Deep Decarbonization Pathways Project (DDPP) is a collaborative initiative to understand and show how individual countries can transition to a low-carbon economy and how the world can meet the internationally agreed target of limiting the increase in global mean surface temperature to less than 2 degrees Celsius (deg. C). Achieving the 2 deg. C limit will require that global net emissions of greenhouse gases (GHG) approach zero by the second half of the century. This will require a profound transformation of energy systems by mid-century through steep declines in carbon intensity in all sectors of the economy, a transition we call 'deep decarbonization.' Successfully transition to a low-carbon economy will require unprecedented global cooperation, including a global cooperative effort to accelerate the development and diffusion of some key low carbon technologies. As underscored throughout this report, the results of the DDPP analyses remain preliminary and incomplete. The DDPP proceeds in two phases. This 2014 report describes the DDPP's approach to deep decarbonization at the country level and presents preliminary findings on technically feasible pathways to deep decarbonization, utilizing technology assumptions and timelines provided by the DDPP Secretariat. At this stage we have not yet considered the economic and social costs and benefits of deep decarbonization, which will be the topic for the next report. The DDPP is issuing this 2014 report to the UN Secretary-General Ban Ki-moon in support of the Climate Leaders' Summit at the United Nations on September 23, 2014. This 2014 report by the Deep Decarbonization Pathway Project (DDPP) summarizes preliminary findings of the technical pathways developed by the DDPP Country Research Partners with the objective of achieving emission reductions consistent with limiting global warming to less than 2 deg. C., without, at this stage, consideration of economic and social costs and benefits. The DDPP is a knowledge

  10. Deep Learning and Bayesian Methods

    Directory of Open Access Journals (Sweden)

    Prosper Harrison B.

    2017-01-01

    Full Text Available A revolution is underway in which deep neural networks are routinely used to solve diffcult problems such as face recognition and natural language understanding. Particle physicists have taken notice and have started to deploy these methods, achieving results that suggest a potentially significant shift in how data might be analyzed in the not too distant future. We discuss a few recent developments in the application of deep neural networks and then indulge in speculation about how such methods might be used to automate certain aspects of data analysis in particle physics. Next, the connection to Bayesian methods is discussed and the paper ends with thoughts on a significant practical issue, namely, how, from a Bayesian perspective, one might optimize the construction of deep neural networks.

  11. Deep Learning in Drug Discovery.

    Science.gov (United States)

    Gawehn, Erik; Hiss, Jan A; Schneider, Gisbert

    2016-01-01

    Artificial neural networks had their first heyday in molecular informatics and drug discovery approximately two decades ago. Currently, we are witnessing renewed interest in adapting advanced neural network architectures for pharmaceutical research by borrowing from the field of "deep learning". Compared with some of the other life sciences, their application in drug discovery is still limited. Here, we provide an overview of this emerging field of molecular informatics, present the basic concepts of prominent deep learning methods and offer motivation to explore these techniques for their usefulness in computer-assisted drug discovery and design. We specifically emphasize deep neural networks, restricted Boltzmann machine networks and convolutional networks. © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.

  12. Eric Davidson and deep time.

    Science.gov (United States)

    Erwin, Douglas H

    2017-10-13

    Eric Davidson had a deep and abiding interest in the role developmental mechanisms played in generating evolutionary patterns documented in deep time, from the origin of the euechinoids to the processes responsible for the morphological architectures of major animal clades. Although not an evolutionary biologist, Davidson's interests long preceded the current excitement over comparative evolutionary developmental biology. Here I discuss three aspects at the intersection between his research and evolutionary patterns in deep time: First, understanding the mechanisms of body plan formation, particularly those associated with the early diversification of major metazoan clades. Second, a critique of early claims about ancestral metazoans based on the discoveries of highly conserved genes across bilaterian animals. Third, Davidson's own involvement in paleontology through a collaborative study of the fossil embryos from the Ediacaran Doushantuo Formation in south China.

  13. Deep Learning in Gastrointestinal Endoscopy.

    Science.gov (United States)

    Patel, Vivek; Armstrong, David; Ganguli, Malika; Roopra, Sandeep; Kantipudi, Neha; Albashir, Siwar; Kamath, Markad V

    2016-01-01

    Gastrointestinal (GI) endoscopy is used to inspect the lumen or interior of the GI tract for several purposes, including, (1) making a clinical diagnosis, in real time, based on the visual appearances; (2) taking targeted tissue samples for subsequent histopathological examination; and (3) in some cases, performing therapeutic interventions targeted at specific lesions. GI endoscopy is therefore predicated on the assumption that the operator-the endoscopist-is able to identify and characterize abnormalities or lesions accurately and reproducibly. However, as in other areas of clinical medicine, such as histopathology and radiology, many studies have documented marked interobserver and intraobserver variability in lesion recognition. Thus, there is a clear need and opportunity for techniques or methodologies that will enhance the quality of lesion recognition and diagnosis and improve the outcomes of GI endoscopy. Deep learning models provide a basis to make better clinical decisions in medical image analysis. Biomedical image segmentation, classification, and registration can be improved with deep learning. Recent evidence suggests that the application of deep learning methods to medical image analysis can contribute significantly to computer-aided diagnosis. Deep learning models are usually considered to be more flexible and provide reliable solutions for image analysis problems compared to conventional computer vision models. The use of fast computers offers the possibility of real-time support that is important for endoscopic diagnosis, which has to be made in real time. Advanced graphics processing units and cloud computing have also favored the use of machine learning, and more particularly, deep learning for patient care. This paper reviews the rapidly evolving literature on the feasibility of applying deep learning algorithms to endoscopic imaging.

  14. Deep mycoses in Amazon region.

    Science.gov (United States)

    Talhari, S; Cunha, M G; Schettini, A P; Talhari, A C

    1988-09-01

    Patients with deep mycoses diagnosed in dermatologic clinics of Manaus (state of Amazonas, Brazil) were studied from November 1973 to December 1983. They came from the Brazilian states of Amazonas, Pará, Acre, and Rondônia and the Federal Territory of Roraima. All of these regions, with the exception of Pará, are situated in the western part of the Amazon Basin. The climatic conditions in this region are almost the same: tropical forest, high rainfall, and mean annual temperature of 26C. The deep mycoses diagnosed, in order of frequency, were Jorge Lobo's disease, paracoccidioidomycosis, chromomycosis, sporotrichosis, mycetoma, cryptococcosis, zygomycosis, and histoplasmosis.

  15. Producing deep-water hydrocarbons

    International Nuclear Information System (INIS)

    Pilenko, Thierry

    2011-01-01

    Several studies relate the history and progress made in offshore production from oil and gas fields in relation to reserves and the techniques for producing oil offshore. The intention herein is not to review these studies but rather to argue that the activities of prospecting and producing deep-water oil and gas call for a combination of technology and project management and, above all, of devotion and innovation. Without this sense of commitment motivating men and women in this industry, the human adventure of deep-water production would never have taken place

  16. Deep soft tissue leiomyoma of the thigh

    Energy Technology Data Exchange (ETDEWEB)

    Watson, G.M.T.; Saifuddin, A. [Department of Radiology, The Royal National Orthopaedic Hospital Trust, Brockley Hill (United Kingdom); Sandison, A. [Department of Pathology, The Royal National Orthopaedic Hospital Trust, Stanmore, Middlesex (United Kingdom)

    1999-07-01

    A case of ossified leiomyoma of the deep soft tissues of the left thigh is presented. The radiographic appearance suggested a low-grade chondrosarcoma. MRI of the lesion showed signal characteristics similar to muscle on both T1- and T2-weighted spin echo sequences with linear areas of high signal intensity on T1-weighted images consistent with medullary fat in metaplastic bone. Histopathological examination of the resected specimen revealed a benign ossified soft tissue leiomyoma. (orig.) With 3 figs., 13 refs.

  17. Deep inelastic scattering and asymptotic freedom

    International Nuclear Information System (INIS)

    Nachtmann, O.

    1985-01-01

    I recall some facets of the history of the field of deep inelastic scattering. I show how there was a very fruitful interplay between phenomenology on the one side and more abstract field theoretical considerations on the other side, where Kurt Symanzik, whose memory we honour today, made important contributions. Finally I make some remarks on the most recent developments in this field which have to do with the so-called EMC-effect, where EMC stands for European Muon Collaboration. (orig./HSI)

  18. Deep inelastic scattering of heavy ions

    International Nuclear Information System (INIS)

    Brink, D.M.

    1980-01-01

    These lecture notes show how path integral methods can be used in the theory of heavy ion reactions. The effects of internal degrees of freedom on the relative motion are contained in an influence functional which is calculated for several simple models of the internal structure. In each model the influence functional has a simple Gaussian structure which suggests that the relative motion of the nuclei in a deep inelastic collision can be described by a Langevin equation. The form of the influence functional determines the average damping force and the correlation function of the fluctuating Langevin force. (author)

  19. Identifying QCD Transition Using Deep Learning

    Science.gov (United States)

    Zhou, Kai; Pang, Long-gang; Su, Nan; Petersen, Hannah; Stoecker, Horst; Wang, Xin-Nian

    2018-02-01

    In this proceeding we review our recent work using supervised learning with a deep convolutional neural network (CNN) to identify the QCD equation of state (EoS) employed in hydrodynamic modeling of heavy-ion collisions given only final-state particle spectra ρ(pT, V). We showed that there is a traceable encoder of the dynamical information from phase structure (EoS) that survives the evolution and exists in the final snapshot, which enables the trained CNN to act as an effective "EoS-meter" in detecting the nature of the QCD transition.

  20. Stimulation Technologies for Deep Well Completions

    Energy Technology Data Exchange (ETDEWEB)

    None

    2003-09-30

    The Department of Energy (DOE) is sponsoring the Deep Trek Program targeted at improving the economics of drilling and completing deep gas wells. Under the DOE program, Pinnacle Technologies is conducting a study to evaluate the stimulation of deep wells. The objective of the project is to assess U.S. deep well drilling & stimulation activity, review rock mechanics & fracture growth in deep, high pressure/temperature wells and evaluate stimulation technology in several key deep plays. An assessment of historical deep gas well drilling activity and forecast of future trends was completed during the first six months of the project; this segment of the project was covered in Technical Project Report No. 1. The second progress report covers the next six months of the project during which efforts were primarily split between summarizing rock mechanics and fracture growth in deep reservoirs and contacting operators about case studies of deep gas well stimulation.

  1. STIMULATION TECHNOLOGIES FOR DEEP WELL COMPLETIONS

    Energy Technology Data Exchange (ETDEWEB)

    Stephen Wolhart

    2003-06-01

    The Department of Energy (DOE) is sponsoring a Deep Trek Program targeted at improving the economics of drilling and completing deep gas wells. Under the DOE program, Pinnacle Technologies is conducting a project to evaluate the stimulation of deep wells. The objective of the project is to assess U.S. deep well drilling & stimulation activity, review rock mechanics & fracture growth in deep, high pressure/temperature wells and evaluate stimulation technology in several key deep plays. Phase 1 was recently completed and consisted of assessing deep gas well drilling activity (1995-2007) and an industry survey on deep gas well stimulation practices by region. Of the 29,000 oil, gas and dry holes drilled in 2002, about 300 were drilled in the deep well; 25% were dry, 50% were high temperature/high pressure completions and 25% were simply deep completions. South Texas has about 30% of these wells, Oklahoma 20%, Gulf of Mexico Shelf 15% and the Gulf Coast about 15%. The Rockies represent only 2% of deep drilling. Of the 60 operators who drill deep and HTHP wells, the top 20 drill almost 80% of the wells. Six operators drill half the U.S. deep wells. Deep drilling peaked at 425 wells in 1998 and fell to 250 in 1999. Drilling is expected to rise through 2004 after which drilling should cycle down as overall drilling declines.

  2. Cognitive Implications of Deep Gray Matter Iron in Multiple Sclerosis.

    Science.gov (United States)

    Fujiwara, E; Kmech, J A; Cobzas, D; Sun, H; Seres, P; Blevins, G; Wilman, A H

    2017-05-01

    Deep gray matter iron accumulation is increasingly recognized in association with multiple sclerosis and can be measured in vivo with MR imaging. The cognitive implications of this pathology are not well-understood, especially vis-à-vis deep gray matter atrophy. Our aim was to investigate the relationships between cognition and deep gray matter iron in MS by using 2 MR imaging-based iron-susceptibility measures. Forty patients with multiple sclerosis (relapsing-remitting, n = 16; progressive, n = 24) and 27 healthy controls were imaged at 4.7T by using the transverse relaxation rate and quantitative susceptibility mapping. The transverse relaxation rate and quantitative susceptibility mapping values and volumes (atrophy) of the caudate, putamen, globus pallidus, and thalamus were determined by multiatlas segmentation. Cognition was assessed with the Brief Repeatable Battery of Neuropsychological Tests. Relationships between cognition and deep gray matter iron were examined by hierarchic regressions. Compared with controls, patients showed reduced memory ( P processing speed ( P = .02) and smaller putamen ( P deep gray matter iron accumulation in the current multiple sclerosis cohort. Atrophy and iron accumulation in deep gray matter both have negative but separable relationships to cognition in multiple sclerosis. © 2017 by American Journal of Neuroradiology.

  3. Predicting healthcare trajectories from medical records: A deep learning approach.

    Science.gov (United States)

    Pham, Trang; Tran, Truyen; Phung, Dinh; Venkatesh, Svetha

    2017-05-01

    Personalized predictive medicine necessitates the modeling of patient illness and care processes, which inherently have long-term temporal dependencies. Healthcare observations, stored in electronic medical records are episodic and irregular in time. We introduce DeepCare, an end-to-end deep dynamic neural network that reads medical records, stores previous illness history, infers current illness states and predicts future medical outcomes. At the data level, DeepCare represents care episodes as vectors and models patient health state trajectories by the memory of historical records. Built on Long Short-Term Memory (LSTM), DeepCare introduces methods to handle irregularly timed events by moderating the forgetting and consolidation of memory. DeepCare also explicitly models medical interventions that change the course of illness and shape future medical risk. Moving up to the health state level, historical and present health states are then aggregated through multiscale temporal pooling, before passing through a neural network that estimates future outcomes. We demonstrate the efficacy of DeepCare for disease progression modeling, intervention recommendation, and future risk prediction. On two important cohorts with heavy social and economic burden - diabetes and mental health - the results show improved prediction accuracy. Copyright © 2017 Elsevier Inc. All rights reserved.

  4. Detecting atrial fibrillation by deep convolutional neural networks.

    Science.gov (United States)

    Xia, Yong; Wulan, Naren; Wang, Kuanquan; Zhang, Henggui

    2018-02-01

    Atrial fibrillation (AF) is the most common cardiac arrhythmia. The incidence of AF increases with age, causing high risks of stroke and increased morbidity and mortality. Efficient and accurate diagnosis of AF based on the ECG is valuable in clinical settings and remains challenging. In this paper, we proposed a novel method with high reliability and accuracy for AF detection via deep learning. The short-term Fourier transform (STFT) and stationary wavelet transform (SWT) were used to analyze ECG segments to obtain two-dimensional (2-D) matrix input suitable for deep convolutional neural networks. Then, two different deep convolutional neural network models corresponding to STFT output and SWT output were developed. Our new method did not require detection of P or R peaks, nor feature designs for classification, in contrast to existing algorithms. Finally, the performances of the two models were evaluated and compared with those of existing algorithms. Our proposed method demonstrated favorable performances on ECG segments as short as 5 s. The deep convolutional neural network using input generated by STFT, presented a sensitivity of 98.34%, specificity of 98.24% and accuracy of 98.29%. For the deep convolutional neural network using input generated by SWT, a sensitivity of 98.79%, specificity of 97.87% and accuracy of 98.63% was achieved. The proposed method using deep convolutional neural networks shows high sensitivity, specificity and accuracy, and, therefore, is a valuable tool for AF detection. Copyright © 2017 Elsevier Ltd. All rights reserved.

  5. Distributed deep learning networks among institutions for medical imaging.

    Science.gov (United States)

    Chang, Ken; Balachandar, Niranjan; Lam, Carson; Yi, Darvin; Brown, James; Beers, Andrew; Rosen, Bruce; Rubin, Daniel L; Kalpathy-Cramer, Jayashree

    2018-03-29

    Deep learning has become a promising approach for automated support for clinical diagnosis. When medical data samples are limited, collaboration among multiple institutions is necessary to achieve high algorithm performance. However, sharing patient data often has limitations due to technical, legal, or ethical concerns. In this study, we propose methods of distributing deep learning models as an attractive alternative to sharing patient data. We simulate the distribution of deep learning models across 4 institutions using various training heuristics and compare the results with a deep learning model trained on centrally hosted patient data. The training heuristics investigated include ensembling single institution models, single weight transfer, and cyclical weight transfer. We evaluated these approaches for image classification in 3 independent image collections (retinal fundus photos, mammography, and ImageNet). We find that cyclical weight transfer resulted in a performance that was comparable to that of centrally hosted patient data. We also found that there is an improvement in the performance of cyclical weight transfer heuristic with a high frequency of weight transfer. We show that distributing deep learning models is an effective alternative to sharing patient data. This finding has implications for any collaborative deep learning study.

  6. Deep Space Climate Observatory (DSCOVR)

    Data.gov (United States)

    National Oceanic and Atmospheric Administration, Department of Commerce — The Deep Space Climate ObserVatoRy (DSCOVR) satellite is a NOAA operated asset at the first Lagrange (L1) point. The primary space weather instrument is the PlasMag...

  7. FOSTERING DEEP LEARNING AMONGST ENTREPRENEURSHIP ...

    African Journals Online (AJOL)

    An important prerequisite for this important objective to be achieved is that lecturers ensure that students adopt a deep learning approach towards entrepreneurship courses been taught, as this will enable them to truly understand key entrepreneurial concepts and strategies and how they can be implemented in the real ...

  8. Deep Space Gateway "Recycler" Mission

    Science.gov (United States)

    Graham, L.; Fries, M.; Hamilton, J.; Landis, R.; John, K.; O'Hara, W.

    2018-02-01

    Use of the Deep Space Gateway provides a hub for a reusable planetary sample return vehicle for missions to gather star dust as well as samples from various parts of the solar system including main belt asteroids, near-Earth asteroids, and Mars moon.

  9. Deep freezers with heat recovery

    Energy Technology Data Exchange (ETDEWEB)

    Kistler, J.

    1981-09-02

    Together with space and water heating systems, deep freezers are the biggest energy consumers in households. The article investigates the possibility of using the waste heat for water heating. The design principle of such a system is presented in a wiring diagram.

  10. A Deep-Sea Simulation.

    Science.gov (United States)

    Montes, Georgia E.

    1997-01-01

    Describes an activity that simulates exploration techniques used in deep-sea explorations and teaches students how this technology can be used to take a closer look inside volcanoes, inspect hazardous waste sites such as nuclear reactors, and explore other environments dangerous to humans. (DDR)

  11. Barbabos Deep-Water Sponges

    NARCIS (Netherlands)

    Soest, van R.W.M.; Stentoft, N.

    1988-01-01

    Deep-water sponges dredged up in two locations off the west coast of Barbados are systematically described. A total of 69 species is recorded, among which 16 are new to science, viz. Pachymatisma geodiformis, Asteropus syringiferus, Cinachyra arenosa, Theonella atlantica. Corallistes paratypus,

  12. Deep learning for visual understanding

    NARCIS (Netherlands)

    Guo, Y.

    2017-01-01

    With the dramatic growth of the image data on the web, there is an increasing demand of the algorithms capable of understanding the visual information automatically. Deep learning, served as one of the most significant breakthroughs, has brought revolutionary success in diverse visual applications,

  13. Deep-Sky Video Astronomy

    CERN Document Server

    Massey, Steve

    2009-01-01

    A guide to using modern integrating video cameras for deep-sky viewing and imaging with the kinds of modest telescopes available commercially to amateur astronomers. It includes an introduction and a brief history of the technology and camera types. It examines the pros and cons of this unrefrigerated yet highly efficient technology

  14. DM Considerations for Deep Drilling

    OpenAIRE

    Dubois-Felsmann, Gregory

    2016-01-01

    An outline of the current situation regarding the DM plans for the Deep Drilling surveys and an invitation to the community to provide feedback on what they would like to see included in the data processing and visualization of these surveys.

  15. Lessons from Earth's Deep Time

    Science.gov (United States)

    Soreghan, G. S.

    2005-01-01

    Earth is a repository of data on climatic changes from its deep-time history. Article discusses the collection and study of these data to predict future climatic changes, the need to create national study centers for the purpose, and the necessary cooperation between different branches of science in climatic research.

  16. Digging Deeper: The Deep Web.

    Science.gov (United States)

    Turner, Laura

    2001-01-01

    Focuses on the Deep Web, defined as Web content in searchable databases of the type that can be found only by direct query. Discusses the problems of indexing; inability to find information not indexed in the search engine's database; and metasearch engines. Describes 10 sites created to access online databases or directly search them. Lists ways…

  17. Deep Learning and Music Adversaries

    DEFF Research Database (Denmark)

    Kereliuk, Corey Mose; Sturm, Bob L.; Larsen, Jan

    2015-01-01

    the minimal perturbation of the input image such that the system misclassifies it with high confidence. We adapt this approach to construct and deploy an adversary of deep learning systems applied to music content analysis. In our case, however, the system inputs are magnitude spectral frames, which require...

  18. Stimulation Technologies for Deep Well Completions

    Energy Technology Data Exchange (ETDEWEB)

    Stephen Wolhart

    2005-06-30

    The Department of Energy (DOE) is sponsoring the Deep Trek Program targeted at improving the economics of drilling and completing deep gas wells. Under the DOE program, Pinnacle Technologies conducted a study to evaluate the stimulation of deep wells. The objective of the project was to review U.S. deep well drilling and stimulation activity, review rock mechanics and fracture growth in deep, high-pressure/temperature wells and evaluate stimulation technology in several key deep plays. This report documents results from this project.

  19. Deep vein thrombosis of the leg

    Energy Technology Data Exchange (ETDEWEB)

    Lee, Eun Hee; Rhee, Kwang Woo; Jeon, Suk Chul; Joo, Kyung Bin; Lee, Seung Ro; Seo, Heung Suk; Hahm, Chang Kok [College of Medicine, Hanyang University, Seoul (Korea, Republic of)

    1987-04-15

    Ascending contrast venography is the definitive standard method for the diagnosis of deep vein thrombosis (DVT) of the lower extremities. Authors analysed 22 cases of DVT clinically and radiographically. 1.The patients ranged in age from 15 to 70 yrs and the most prevalent age group was 7th decade (31%). There was an equal distribution of males and females. 2.In 11 cases of 22 cases, variable etiologic factors were recognized, such as abdominal surgery, chronic bedridden state, local trauma on the leg, pregnancy, postpartum, Behcet's syndrome, iliac artery aneurysm, and chronic medication of estrogen. 3.Nineteen cases out of 22 cases showed primary venographic signs of DVT, such as well-defined filling defect in opacified veins and narrowed, irregularly filled venous lumen. In only 3 cases, the diagnosis of DVT was base upon the segmental nonvisualization of deep veins with good opacification of proximal and distal veins and presence of collaterals. 4.Extent of thrombosis: 3 cases were confined to calf vein, 4 cases extended to femoral vein, and 15 cases had involvement above iliac vein. 5.In 17 cases involving relatively long segment of deep veins, propagation pattern of thrombus was evaluated by its radiologic morphology according to the age of thrombus: 9 cases suggested central or antegrade propagation pattern and 8 cases, peripheral or retrograde pattern. 6.None of 22 cases showed clinical evidence of pulmonary embolism. The cause of the rarity of pulmonary embolism in Korean in presumed to be related to the difference in major involving site and propagation pattern of DVT in the leg.

  20. Deep vein thrombosis of the leg

    International Nuclear Information System (INIS)

    Lee, Eun Hee; Rhee, Kwang Woo; Jeon, Suk Chul; Joo, Kyung Bin; Lee, Seung Ro; Seo, Heung Suk; Hahm, Chang Kok

    1987-01-01

    Ascending contrast venography is the definitive standard method for the diagnosis of deep vein thrombosis (DVT) of the lower extremities. Authors analysed 22 cases of DVT clinically and radiographically. 1.The patients ranged in age from 15 to 70 yrs and the most prevalent age group was 7th decade (31%). There was an equal distribution of males and females. 2.In 11 cases of 22 cases, variable etiologic factors were recognized, such as abdominal surgery, chronic bedridden state, local trauma on the leg, pregnancy, postpartum, Behcet's syndrome, iliac artery aneurysm, and chronic medication of estrogen. 3.Nineteen cases out of 22 cases showed primary venographic signs of DVT, such as well-defined filling defect in opacified veins and narrowed, irregularly filled venous lumen. In only 3 cases, the diagnosis of DVT was base upon the segmental nonvisualization of deep veins with good opacification of proximal and distal veins and presence of collaterals. 4.Extent of thrombosis: 3 cases were confined to calf vein, 4 cases extended to femoral vein, and 15 cases had involvement above iliac vein. 5.In 17 cases involving relatively long segment of deep veins, propagation pattern of thrombus was evaluated by its radiologic morphology according to the age of thrombus: 9 cases suggested central or antegrade propagation pattern and 8 cases, peripheral or retrograde pattern. 6.None of 22 cases showed clinical evidence of pulmonary embolism. The cause of the rarity of pulmonary embolism in Korean in presumed to be related to the difference in major involving site and propagation pattern of DVT in the leg

  1. UV Photography Shows Hidden Sun Damage

    Science.gov (United States)

    ... mcat1=de12", ]; for (var c = 0; c UV photography shows hidden sun damage A UV photograph gives ... developing skin cancer and prematurely aged skin. Normal photography UV photography 18 months of age: This boy's ...

  2. Quasi-periodicity in deep redshift surveys

    International Nuclear Information System (INIS)

    Weygaert, R. van de

    1991-01-01

    The recent result by Broadhurst et al., (1990. Nature 343, 726) showing a striking, nearly periodic, galaxy redshift distribution in a narrow pencil-beam survey, is explained within the Voronoi cellular model of clustering of galaxies. Galaxies, whose luminosities are selected from a Schechter luminosity function, are placed randomly within the walls of this cellular model. Narrow and deep, magnitude-limited, pencil-beam surveys through these structures are simulated. Some 15 per cent of these beams show that observed regular pattern, with a spacing between the peaks of the order of 105 h -1 -150 h -1 Mpc, but most pencil-beams show peaks in the redshift distribution without periodicity, so we may conclude that, even within a cellular universe, periodicity is not a common phenomenon. (author)

  3. Periodic arrays of deep nanopores made in silicon with reactive ion etching and deep UV lithography

    International Nuclear Information System (INIS)

    Woldering, Leon A; Tjerkstra, R Willem; Vos, Willem L; Jansen, Henri V; Setija, Irwan D

    2008-01-01

    We report on the fabrication of periodic arrays of deep nanopores with high aspect ratios in crystalline silicon. The radii and pitches of the pores were defined in a chromium mask by means of deep UV scan and step technology. The pores were etched with a reactive ion etching process with SF 6 , optimized for the formation of deep nanopores. We have realized structures with pitches between 440 and 750 nm, pore diameters between 310 and 515 nm, and depth to diameter aspect ratios up to 16. To the best of our knowledge, this is the highest aspect ratio ever reported for arrays of nanopores in silicon made with a reactive ion etching process. Our experimental results show that the etching rate of the nanopores is aspect-ratio-dependent, and is mostly influenced by the angular distribution of the etching ions. Furthermore we show both experimentally and theoretically that, for sub-micrometer structures, reducing the sidewall erosion is the best way to maximize the aspect ratio of the pores. Our structures have potential applications in chemical sensors, in the control of liquid wetting of surfaces, and as capacitors in high-frequency electronics. We demonstrate by means of optical reflectivity that our high-quality structures are very well suited as photonic crystals. Since the process studied is compatible with existing CMOS semiconductor fabrication, it allows for the incorporation of the etched arrays in silicon chips

  4. Effect of sevoflurane on neuronal activity during deep brain stimulation surgery for epilepsy: A case report

    OpenAIRE

    Michaël J. Bos, MD; Linda Ackermans, MD, PhD; Frédéric L.W.V.J. Schaper, MD; Rob P.W. Rouhl, MD, PhD; Vivianne H.J.M. van Kranen-Mastenbroek, MD, PhD; Wolfgang F. Buhre, MD, PhD; Marcus L.F. Janssen, MD, PhD

    2018-01-01

    Deep brain stimulation of the anterior nucleus of the thalamus is an effective treatment for patients with refractory epilepsy who do not respond sufficiently to medical therapy. Optimal therapeutic effects of deep brain stimulation probably depend on accurate positioning of the stimulating electrodes. Microelectrode recordings show bursty firing neurons in the anterior nucleus of the thalamus region, which confirms the anatomical target determined by the surgeon. Deep brain stimulation elect...

  5. Characterization of failure modes in deep UV and deep green LEDs utilizing advanced semiconductor localization techniques.

    Energy Technology Data Exchange (ETDEWEB)

    Tangyunyong, Paiboon; Miller, Mary A.; Cole, Edward Isaac, Jr.

    2012-03-01

    We present the results of a two-year early career LDRD that focused on defect localization in deep green and deep ultraviolet (UV) light-emitting diodes (LEDs). We describe the laser-based techniques (TIVA/LIVA) used to localize the defects and interpret data acquired. We also describe a defect screening method based on a quick electrical measurement to determine whether defects should be present in the LEDs. We then describe the stress conditions that caused the devices to fail and how the TIVA/LIVA techniques were used to monitor the defect signals as the devices degraded and failed. We also describe the correlation between the initial defects and final degraded or failed state of the devices. Finally we show characterization results of the devices in the failed conditions and present preliminary theories as to why the devices failed for both the InGaN (green) and AlGaN (UV) LEDs.

  6. Educational Outreach: The Space Science Road Show

    Science.gov (United States)

    Cox, N. L. J.

    2002-01-01

    The poster presented will give an overview of a study towards a "Space Road Show". The topic of this show is space science. The target group is adolescents, aged 12 to 15, at Dutch high schools. The show and its accompanying experiments would be supported with suitable educational material. Science teachers at schools can decide for themselves if they want to use this material in advance, afterwards or not at all. The aims of this outreach effort are: to motivate students for space science and engineering, to help them understand the importance of (space) research, to give them a positive feeling about the possibilities offered by space and in the process give them useful knowledge on space basics. The show revolves around three main themes: applications, science and society. First the students will get some historical background on the importance of space/astronomy to civilization. Secondly they will learn more about novel uses of space. On the one hand they will learn of "Views on Earth" involving technologies like Remote Sensing (or Spying), Communication, Broadcasting, GPS and Telemedicine. On the other hand they will experience "Views on Space" illustrated by past, present and future space research missions, like the space exploration missions (Cassini/Huygens, Mars Express and Rosetta) and the astronomy missions (Soho and XMM). Meanwhile, the students will learn more about the technology of launchers and satellites needed to accomplish these space missions. Throughout the show and especially towards the end attention will be paid to the third theme "Why go to space"? Other reasons for people to get into space will be explored. An important question in this is the commercial (manned) exploration of space. Thus, the questions of benefit of space to society are integrated in the entire show. It raises some fundamental questions about the effects of space travel on our environment, poverty and other moral issues. The show attempts to connect scientific with

  7. 2008 LHC Open Days Physics: the show

    CERN Multimedia

    2008-01-01

    A host of events and activities await visitors to the LHC Open Days on 5 and 6 April. A highlight will be the physics shows funded by the European Physical Society (EPS), which are set to surprise and challenge children and adults alike! School children use their experience of riding a bicycle to understand how planets move around the sun (Copyright : Circus Naturally) Participating in the Circus Naturally show could leave a strange taste in your mouth! (Copyright : Circus Naturally) The Rino Foundation’s experiments with liquid nitrogen can be pretty exciting! (Copyright: The Rino Foundation)What does a bicycle have in common with the solar system? Have you ever tried to weigh air or visualise sound? Ever heard of a vacuum bazooka? If you want to discover the answers to these questions and more then come to the Physics Shows taking place at the CERN O...

  8. Online Italian fandoms of American TV shows

    Directory of Open Access Journals (Sweden)

    Eleonora Benecchi

    2015-06-01

    Full Text Available The Internet has changed media fandom in two main ways: it helps fans connect with each other despite physical distance, leading to the formation of international fan communities; and it helps fans connect with the creators of the TV show, deepening the relationship between TV producers and international fandoms. To assess whether Italian fan communities active online are indeed part of transnational online communities and whether the Internet has actually altered their relationship with the creators of the original text they are devoted to, qualitative analysis and narrative interviews of 26 Italian fans of American TV shows were conducted to explore the fan-producer relationship. Results indicated that the online Italian fans surveyed preferred to stay local, rather than using geography-leveling online tools. Further, the sampled Italian fans' relationships with the show runners were mediated or even absent.

  9. A new data transmission system for deep water applications

    International Nuclear Information System (INIS)

    Brown, Gerald K.

    2000-01-01

    A novel data transmission system is now available. Conventional data transmission methods include systems that require satellites, hard wires, fiber optics and other methods that do not lend themselves to buried, remote, or deep water applications. The Data Transmission System (DTS) induces a signal into a structure such as the transmission line and retrieving the signal at a distant point. In deep water applications the power required comes from an anode array that generates its own power. In addition to deep water applications, the DTS can be used in onshore, drilling, and downhole applications. With repeater stations, most lengths of gathering and transmission lines can be used. Therefore data from control valves, strain gauges, corrosion monitoring, sand monitoring, valve position and other process variables can all be transmitted. Comparisons are made between the different data transmission systems showing the advantages and disadvantages of each type with comparative costs showing the advantages of the new DTS system. (author)

  10. Deep Web and Dark Web: Deep World of the Internet

    OpenAIRE

    Çelik, Emine

    2018-01-01

    The Internet is undoubtedly still a revolutionary breakthrough in the history of humanity. Many people use the internet for communication, social media, shopping, political and social agenda, and more. Deep Web and Dark Web concepts not only handled by computer, software engineers but also handled by social siciensists because of the role of internet for the States in international arenas, public institutions and human life. By the moving point that very importantrole of internet for social s...

  11. Deep Phenotyping: Deep Learning For Temporal Phenotype/Genotype Classification

    OpenAIRE

    Najafi, Mohammad; Namin, Sarah; Esmaeilzadeh, Mohammad; Brown, Tim; Borevitz, Justin

    2017-01-01

    High resolution and high throughput, genotype to phenotype studies in plants are underway to accelerate breeding of climate ready crops. Complex developmental phenotypes are observed by imaging a variety of accessions in different environment conditions, however extracting the genetically heritable traits is challenging. In the recent years, deep learning techniques and in particular Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs) and Long-Short Term Memories (LSTMs), h...

  12. Deep Neuromuscular Blockade Improves Laparoscopic Surgical Conditions

    DEFF Research Database (Denmark)

    Rosenberg, Jacob; Herring, W Joseph; Blobner, Manfred

    2017-01-01

    INTRODUCTION: Sustained deep neuromuscular blockade (NMB) during laparoscopic surgery may facilitate optimal surgical conditions. This exploratory study assessed whether deep NMB improves surgical conditions and, in doing so, allows use of lower insufflation pressures during laparoscopic cholecys...

  13. Joint Training of Deep Boltzmann Machines

    OpenAIRE

    Goodfellow, Ian; Courville, Aaron; Bengio, Yoshua

    2012-01-01

    We introduce a new method for training deep Boltzmann machines jointly. Prior methods require an initial learning pass that trains the deep Boltzmann machine greedily, one layer at a time, or do not perform well on classifi- cation tasks.

  14. Duchenne muscular dystrophy models show their age

    OpenAIRE

    Chamberlain, Jeffrey S.

    2010-01-01

    The lack of appropriate animal models has hampered efforts to develop therapies for Duchenne muscular dystrophy (DMD). A new mouse model lacking both dystrophin and telomerase (Sacco et al., 2010) closely mimics the pathological progression of human DMD and shows that muscle stem cell activity is a key determinant of disease severity.

  15. Show Them You Really Want the Job

    Science.gov (United States)

    Perlmutter, David D.

    2012-01-01

    Showing that one really "wants" the job entails more than just really wanting the job. An interview is part Broadway casting call, part intellectual dating game, part personality test, and part, well, job interview. When there are 300 applicants for a position, many of them will "fit" the required (and even the preferred) skills listed in the job…

  16. A Talk Show from the Past.

    Science.gov (United States)

    Gallagher, Arlene F.

    1991-01-01

    Describes a two-day activity in which elementary students examine voting rights, the right to assemble, and women's suffrage. Explains the game, "Assemble, Reassemble," and a student-produced talk show with five students playing the roles of leaders of the women's suffrage movement. Profiles Elizabeth Cady Stanton, Lucretia Mott, Susan…

  17. Laser entertainment and light shows in education

    Science.gov (United States)

    Sabaratnam, Andrew T.; Symons, Charles

    2002-05-01

    Laser shows and beam effects have been a source of entertainment since its first public performance May 9, 1969, at Mills College in Oakland, California. Since 1997, the Photonics Center, NgeeAnn Polytechnic, Singapore, has been using laser shows as a teaching tool. Students are able to exhibit their creative skills and learn at the same time how lasers are used in the entertainment industry. Students will acquire a number of skills including handling three- phase power supply, operation of cooling system, and laser alignment. Students also acquire an appreciation of the arts, learning about shapes and contours as they develop graphics for the shows. After holography, laser show animation provides a combination of the arts and technology. This paper aims to briefly describe how a krypton-argon laser, galvanometer scanners, a polychromatic acousto-optic modulator and related electronics are put together to develop a laser projector. The paper also describes how students are trained to make their own laser animation and beam effects with music, and at the same time have an appreciation of the operation of a Class IV laser and the handling of optical components.

  18. The Last Great American Picture Show

    NARCIS (Netherlands)

    Elsaesser, Thomas; King, Noel; Horwath, Alexander

    2004-01-01

    The Last Great American Picture Show brings together essays by scholars and writers who chart the changing evaluations of the American cinema of the 1970s, sometimes referred to as the decade of the lost generation, but now more and more recognized as the first New Hollywood, without which the

  19. Deep and intermediate mediterranean water in the western Alboran Sea

    Science.gov (United States)

    Parrilla, Gregorio; Kinder, Thomas H.; Preller, Ruth H.

    1986-01-01

    Hydrographic and current meter data, obtained during June to October 1982, and numerical model experiments are used to study the distribution and flow of Mediterranean waters in the western Alboran Sea. The Intermediate Water is more pronounced in the northern three-fourths of the sea, but its distribution is patchy as manifested by variability of the temperature and salinity maxima at scales ≤10 km. Current meters in the lower Intermediate Water showed mean flow toward the Strait at 2 cm s -1. A reversal of this flow lasted about 2 weeks. A rough estimate of the mean westward Intermediate Water transport was 0.4 × 10 6 m 3 s -1, about one-third of the total outflow, so that the best estimates of the contributions of traditionally defined Intermediate Water and Deep Water account for only about one-half of the total outflow. The Deep Water was uplifted against the southern continental slope from Alboran Island (3°W) to the Strait. There was also a similar but much weaker banking against the Spanish slope, but a deep current record showed that the eastward recirculation implied by this banking is probably intermittent. Two-layer numerical model experiments simulated the Intermediate Water flow with a flat bottom and the Deep Water with realistic bottom topography. Both experiments replicated the major circulation features, and the Intermediate Water flow was concentrated in the north because of rotation and the Deep Water flow in the south because of topographic control.

  20. Automatic activation of word phonology from print in deep dyslexia.

    Science.gov (United States)

    Katz, R B; Lanzoni, S M

    1992-11-01

    The performance of deep dyslexics in oral reading and other tasks suggests that they are poor at activating the phonology of words and non-words from printed stimuli. As the tasks ordinarily used to test deep dyslexics require controlled processing, it is possible that the phonology of printed words can be better activated on an automatic basis. This study investigated this possibility by testing a deep dyslexic patient on a lexical decision task with pairs of stimuli presented simultaneously. In Experiment 1, which used content words as stimuli, the deep dyslexic, like normal subjects, showed faster reaction times on trials with rhyming, similarly spelled stimuli (e.g. bribe-tribe) than on control trials (consisting of non-rhyming, dissimilarly spelled words), but slower reaction times on trials with non-rhyming, similarly spelled stimuli (e.g. couch-touch). When the experiment was repeated using function words as stimuli, the patient no longer showed a phonological effect. Therefore, the phonological activation of printed content words by deep dyslexics may be better than would be expected on the basis of their oral reading performance.

  1. Building Program Vector Representations for Deep Learning

    OpenAIRE

    Mou, Lili; Li, Ge; Liu, Yuxuan; Peng, Hao; Jin, Zhi; Xu, Yan; Zhang, Lu

    2014-01-01

    Deep learning has made significant breakthroughs in various fields of artificial intelligence. Advantages of deep learning include the ability to capture highly complicated features, weak involvement of human engineering, etc. However, it is still virtually impossible to use deep learning to analyze programs since deep architectures cannot be trained effectively with pure back propagation. In this pioneering paper, we propose the "coding criterion" to build program vector representations, whi...

  2. Deep Packet/Flow Analysis using GPUs

    Energy Technology Data Exchange (ETDEWEB)

    Gong, Qian [Fermi National Accelerator Lab. (FNAL), Batavia, IL (United States); Wu, Wenji [Fermi National Accelerator Lab. (FNAL), Batavia, IL (United States); DeMar, Phil [Fermi National Accelerator Lab. (FNAL), Batavia, IL (United States)

    2017-11-12

    Deep packet inspection (DPI) faces severe performance challenges in high-speed networks (40/100 GE) as it requires a large amount of raw computing power and high I/O throughputs. Recently, researchers have tentatively used GPUs to address the above issues and boost the performance of DPI. Typically, DPI applications involve highly complex operations in both per-packet and per-flow data level, often in real-time. The parallel architecture of GPUs fits exceptionally well for per-packet network traffic processing. However, for stateful network protocols such as TCP, their data stream need to be reconstructed in a per-flow level to deliver a consistent content analysis. Since the flow-centric operations are naturally antiparallel and often require large memory space for buffering out-of-sequence packets, they can be problematic for GPUs, whose memory is normally limited to several gigabytes. In this work, we present a highly efficient GPU-based deep packet/flow analysis framework. The proposed design includes a purely GPU-implemented flow tracking and TCP stream reassembly. Instead of buffering and waiting for TCP packets to become in sequence, our framework process the packets in batch and uses a deterministic finite automaton (DFA) with prefix-/suffix- tree method to detect patterns across out-of-sequence packets that happen to be located in different batches. In conclusion, evaluation shows that our code can reassemble and forward tens of millions of packets per second and conduct a stateful signature-based deep packet inspection at 55 Gbit/s using an NVIDIA K40 GPU.

  3. Generating Seismograms with Deep Neural Networks

    Science.gov (United States)

    Krischer, L.; Fichtner, A.

    2017-12-01

    The recent surge of successful uses of deep neural networks in computer vision, speech recognition, and natural language processing, mainly enabled by the availability of fast GPUs and extremely large data sets, is starting to see many applications across all natural sciences. In seismology these are largely confined to classification and discrimination tasks. In this contribution we explore the use of deep neural networks for another class of problems: so called generative models.Generative modelling is a branch of statistics concerned with generating new observed data samples, usually by drawing from some underlying probability distribution. Samples with specific attributes can be generated by conditioning on input variables. In this work we condition on seismic source (mechanism and location) and receiver (location) parameters to generate multi-component seismograms.The deep neural networks are trained on synthetic data calculated with Instaseis (http://instaseis.net, van Driel et al. (2015)) and waveforms from the global ShakeMovie project (http://global.shakemovie.princeton.edu, Tromp et al. (2010)). The underlying radially symmetric or smoothly three dimensional Earth structures result in comparatively small waveform differences from similar events or at close receivers and the networks learn to interpolate between training data samples.Of particular importance is the chosen misfit functional. Generative adversarial networks (Goodfellow et al. (2014)) implement a system in which two networks compete: the generator network creates samples and the discriminator network distinguishes these from the true training examples. Both are trained in an adversarial fashion until the discriminator can no longer distinguish between generated and real samples. We show how this can be applied to seismograms and in particular how it compares to networks trained with more conventional misfit metrics. Last but not least we attempt to shed some light on the black-box nature of

  4. Is Multitask Deep Learning Practical for Pharma?

    Science.gov (United States)

    Ramsundar, Bharath; Liu, Bowen; Wu, Zhenqin; Verras, Andreas; Tudor, Matthew; Sheridan, Robert P; Pande, Vijay

    2017-08-28

    Multitask deep learning has emerged as a powerful tool for computational drug discovery. However, despite a number of preliminary studies, multitask deep networks have yet to be widely deployed in the pharmaceutical and biotech industries. This lack of acceptance stems from both software difficulties and lack of understanding of the robustness of multitask deep networks. Our work aims to resolve both of these barriers to adoption. We introduce a high-quality open-source implementation of multitask deep networks as part of the DeepChem open-source platform. Our implementation enables simple python scripts to construct, fit, and evaluate sophisticated deep models. We use our implementation to analyze the performance of multitask deep networks and related deep models on four collections of pharmaceutical data (three of which have not previously been analyzed in the literature). We split these data sets into train/valid/test using time and neighbor splits to test multitask deep learning performance under challenging conditions. Our results demonstrate that multitask deep networks are surprisingly robust and can offer strong improvement over random forests. Our analysis and open-source implementation in DeepChem provide an argument that multitask deep networks are ready for widespread use in commercial drug discovery.

  5. Evaluation of the DeepWind concept

    DEFF Research Database (Denmark)

    Schmidt Paulsen, Uwe; Borg, Michael; Gonzales Seabra, Luis Alberto

    The report describes the DeepWind 5 MW conceptual design as a baseline for results obtained in the scientific and technical work packages of the DeepWind project. A comparison of DeepWi nd with existing VAWTs and paper projects are carried out and the evaluation of the concept in terms of cost...

  6. Consolidated Deep Actor Critic Networks (DRAFT)

    NARCIS (Netherlands)

    Van der Laan, T.A.

    2015-01-01

    The works [Volodymyr et al. Playing atari with deep reinforcement learning. arXiv preprint arXiv:1312.5602, 2013.] and [Volodymyr et al. Human-level control through deep reinforcement learning. Nature, 518(7540):529–533, 2015.] have demonstrated the power of combining deep neural networks with

  7. Simulator Studies of the Deep Stall

    Science.gov (United States)

    White, Maurice D.; Cooper, George E.

    1965-01-01

    Simulator studies of the deep-stall problem encountered with modern airplanes are discussed. The results indicate that the basic deep-stall tendencies produced by aerodynamic characteristics are augmented by operational considerations. Because of control difficulties to be anticipated in the deep stall, it is desirable that adequate safeguards be provided against inadvertent penetrations.

  8. TOPIC MODELING: CLUSTERING OF DEEP WEBPAGES

    OpenAIRE

    Muhunthaadithya C; Rohit J.V; Sadhana Kesavan; E. Sivasankar

    2015-01-01

    The internet is comprised of massive amount of information in the form of zillions of web pages.This information can be categorized into the surface web and the deep web. The existing search engines can effectively make use of surface web information.But the deep web remains unexploited yet. Machine learning techniques have been commonly employed to access deep web content.

  9. Reality, ficción o show

    Directory of Open Access Journals (Sweden)

    Sandra Ruíz Moreno

    2002-01-01

    Full Text Available Para tener un punto de vista claro y objetivo frente a la polémica establecida en torno al programa “Protagonistas de novela” y la tendiente proliferación de los reality show en las parrillas de programación de la televisión colombiana, se realizó un análisis de texto y contenido de dicho programa, intentando definirlo desde sus posibilidades de realidad, ficción y show. Las unidades de análisis y el estudio de su tratamiento arrojaron un alto contenido que gira en torno a las emociones del ser humano relacionadas con la convivencia, tratadas a manera de show y con algunos aportes textuales de ficción, pero sin su elemento mediador básico, el actor, quitándole toda la posibilidad de tener un tratamiento con la profundidad, distancia y ética que requieren los temas de esta índole. El resultado es un formato que sólo busca altos índices de sintonía y que pertenece más a la denominada televisión “trash”, que a una búsqueda de realidad del hombre y mucho menos de sociedad.

  10. DeepFlavour in CMS

    CERN Multimedia

    CERN. Geneva

    2017-01-01

    Flavour-tagging of jets is an important task in collider based high energy physics and a field where machine learning tools are applied by all major experiments. A new tagger (DeepFlavour) was developed and commissioned in CMS that is based on an advanced machine learning procedure. A deep neural network is used to do multi-classification of jets that origin from a b-quark, two b-quarks, a c-quark, two c-quarks or light colored particles (u, d, s-quark or gluon). The performance was measured in both, data and simulation. The talk will also include the measured performance of all taggers in CMS. The different taggers and results will be discussed and compared with some focus on details of the newest tagger.

  11. Deep Space Habitat Concept Demonstrator

    Science.gov (United States)

    Bookout, Paul S.; Smitherman, David

    2015-01-01

    This project will develop, integrate, test, and evaluate Habitation Systems that will be utilized as technology testbeds and will advance NASA's understanding of alternative deep space mission architectures, requirements, and operations concepts. Rapid prototyping and existing hardware will be utilized to develop full-scale habitat demonstrators. FY 2014 focused on the development of a large volume Space Launch System (SLS) class habitat (Skylab Gen 2) based on the SLS hydrogen tank components. Similar to the original Skylab, a tank section of the SLS rocket can be outfitted with a deep space habitat configuration and launched as a payload on an SLS rocket. This concept can be used to support extended stay at the Lunar Distant Retrograde Orbit to support the Asteroid Retrieval Mission and provide a habitat suitable for human missions to Mars.

  12. Hybrid mask for deep etching

    KAUST Repository

    Ghoneim, Mohamed T.

    2017-08-10

    Deep reactive ion etching is essential for creating high aspect ratio micro-structures for microelectromechanical systems, sensors and actuators, and emerging flexible electronics. A novel hybrid dual soft/hard mask bilayer may be deposited during semiconductor manufacturing for deep reactive etches. Such a manufacturing process may include depositing a first mask material on a substrate; depositing a second mask material on the first mask material; depositing a third mask material on the second mask material; patterning the third mask material with a pattern corresponding to one or more trenches for transfer to the substrate; transferring the pattern from the third mask material to the second mask material; transferring the pattern from the second mask material to the first mask material; and/or transferring the pattern from the first mask material to the substrate.

  13. Map showing sediment isopachs in the deep-sea basins of the Pacific continental margin, Point Conception to Point Loma

    Science.gov (United States)

    Gardner, J.V.; Cacchione, D.A.; Drake, D.E.; Edwards, B.D.; Field, M.E.; Hampton, M.A.; Karl, Herman A.; Kenyon, Neil H.; Masson, D.G.; McCulloch, D.S.; Grim, M.S.

    1992-01-01

    The U.S. Geological Survey conducted a series of cruises, EEZSCAN 84 (EEZ-SCAN 84 Scientific Staff, 1986), to collect reconnaissance data on the newly proclaimed Exclusive Economic Zone (EEZ), the area out to 200 nautical miles from the coastline of the United States. The cruises systematically surveyed the entire conterminous United States west coast EEZ using the Geological Long-Range Inclined Asdic (GLORIA) side-scan sonar, a 160-in3 airgun seismic-reflection profiler, a 3.5-kHz high-resolution seismic-reflection profiler, a 10-kHz echo sounder, and a proton-precession magnetometer. The nominal trackline spacing throughout the survey was 30 km.

  14. Map showing sediment isopachs in the deep-sea basins of the Pacific continental margin, Cape Mendocino to Point Conception

    Science.gov (United States)

    Gardner, J.V.; Cacchione, D.A.; Drake, D.E.; Edwards, B.D.; Field, M.E.; Hampton, M.A.; Karl, Herman A.; Kenyon, Neil H.; Masson, D.G.; McCulloch, D.S.; Grim, M.S.

    1993-01-01

    The U.S. Geological Survey conducted a series of cruises, EEZSCAN 84 (EEZ-SCAN 84 Scientific Staff, 1986), to collect reconnaissance data on the newly proclaimed Exclusive Economic Zone (EEZ), the area out to 200 nautical miles from the coastline of the United States. The cruises systematically surveyed the entire conterminous United States west coast EEZ using the Geological Long-Range Inclined Asdic (GLORIA) side-scan sonar, a 160-in3 airgun seismic-reflection profiler, a 3.5-kHz high-resolution seismic-reflection profiler, a 10-kHz echo sounder, and a proton-precession magnetometer. The nominal trackline spacing throughout the survey was 30 km.

  15. Soft-Deep Boltzmann Machines

    OpenAIRE

    Kiwaki, Taichi

    2015-01-01

    We present a layered Boltzmann machine (BM) that can better exploit the advantages of a distributed representation. It is widely believed that deep BMs (DBMs) have far greater representational power than its shallow counterpart, restricted Boltzmann machines (RBMs). However, this expectation on the supremacy of DBMs over RBMs has not ever been validated in a theoretical fashion. In this paper, we provide both theoretical and empirical evidences that the representational power of DBMs can be a...

  16. Bleeding during gonioscopy after deep sclerectomy.

    Science.gov (United States)

    Moreno-Montañés, Javier; Rodríguez-Conde, Rosa

    2003-10-01

    To show a new complication after deep sclerectomy (DS). We described two eyes of two patients with open-angle glaucoma and cataract who were operated on of an uneventful phacoemulsification and DS with SK-gel implantation. Bleeding during gonioscopic examination occurred in both eyes 7 and 8 months after combined surgery. The blood originated from the vessels around the Descemet window, and was probably due to manipulation or rocking of the goniolens. Pressure was immediately applied to the gonioscopic lens and the hyphema was interrupted. These cases show the presence of new vessels around the Descemet window after DS with SK-gel. Bleeding from the Descemet window vessels can occur during gonioscopy even months after DS. We recommend conducting a careful gonioscopic examination in patients who have undergone DS to avoid this complication.

  17. Deep-learnt classification of light curves

    DEFF Research Database (Denmark)

    Mahabal, Ashish; Gieseke, Fabian; Pai, Akshay Sadananda Uppinakudru

    2017-01-01

    is to derive statistical features from the time series and to use machine learning methods, generally supervised, to separate objects into a few of the standard classes. In this work, we transform the time series to two-dimensional light curve representations in order to classify them using modern deep......Astronomy light curves are sparse, gappy, and heteroscedastic. As a result standard time series methods regularly used for financial and similar datasets are of little help and astronomers are usually left to their own instruments and techniques to classify light curves. A common approach...... learning techniques. In particular, we show that convolutional neural networks based classifiers work well for broad characterization and classification. We use labeled datasets of periodic variables from CRTS survey and show how this opens doors for a quick classification of diverse classes with several...

  18. Deep Learning Improves Antimicrobial Peptide Recognition.

    Science.gov (United States)

    Veltri, Daniel; Kamath, Uday; Shehu, Amarda

    2018-03-24

    Bacterial resistance to antibiotics is a growing concern. Antimicrobial peptides (AMPs), natural components of innate immunity, are popular targets for developing new drugs. Machine learning methods are now commonly adopted by wet-laboratory researchers to screen for promising candidates. In this work we utilize deep learning to recognize antimicrobial activity. We propose a neural network model with convolutional and recurrent layers that leverage primary sequence composition. Results show that the proposed model outperforms state-of-the-art classification models on a comprehensive data set. By utilizing the embedding weights, we also present a reduced-alphabet representation and show that reasonable AMP recognition can be maintained using nine amino-acid types. Models and data sets are made freely available through the Antimicrobial Peptide Scanner vr.2 web server at: www.ampscanner.com. amarda@gmu.edu for general inquiries and dan.veltri@gmail.com for web server information. Supplementary data are available at Bioinformatics online.

  19. Evolving Deep Networks Using HPC

    Energy Technology Data Exchange (ETDEWEB)

    Young, Steven R. [ORNL, Oak Ridge; Rose, Derek C. [ORNL, Oak Ridge; Johnston, Travis [ORNL, Oak Ridge; Heller, William T. [ORNL, Oak Ridge; Karnowski, thomas P. [ORNL, Oak Ridge; Potok, Thomas E. [ORNL, Oak Ridge; Patton, Robert M. [ORNL, Oak Ridge; Perdue, Gabriel [Fermilab; Miller, Jonathan [Santa Maria U., Valparaiso

    2017-01-01

    While a large number of deep learning networks have been studied and published that produce outstanding results on natural image datasets, these datasets only make up a fraction of those to which deep learning can be applied. These datasets include text data, audio data, and arrays of sensors that have very different characteristics than natural images. As these “best” networks for natural images have been largely discovered through experimentation and cannot be proven optimal on some theoretical basis, there is no reason to believe that they are the optimal network for these drastically different datasets. Hyperparameter search is thus often a very important process when applying deep learning to a new problem. In this work we present an evolutionary approach to searching the possible space of network hyperparameters and construction that can scale to 18, 000 nodes. This approach is applied to datasets of varying types and characteristics where we demonstrate the ability to rapidly find best hyperparameters in order to enable practitioners to quickly iterate between idea and result.

  20. Deep Space Gateway Science Opportunities

    Science.gov (United States)

    Quincy, C. D.; Charles, J. B.; Hamill, Doris; Sidney, S. C.

    2018-01-01

    The NASA Life Sciences Research Capabilities Team (LSRCT) has been discussing deep space research needs for the last two years. NASA's programs conducting life sciences studies - the Human Research Program, Space Biology, Astrobiology, and Planetary Protection - see the Deep Space Gateway (DSG) as affording enormous opportunities to investigate biological organisms in a unique environment that cannot be replicated in Earth-based laboratories or on Low Earth Orbit science platforms. These investigations may provide in many cases the definitive answers to risks associated with exploration and living outside Earth's protective magnetic field. Unlike Low Earth Orbit or terrestrial locations, the Gateway location will be subjected to the true deep space spectrum and influence of both galactic cosmic and solar particle radiation and thus presents an opportunity to investigate their long-term exposure effects. The question of how a community of biological organisms change over time within the harsh environment of space flight outside of the magnetic field protection can be investigated. The biological response to the absence of Earth's geomagnetic field can be studied for the first time. Will organisms change in new and unique ways under these new conditions? This may be specifically true on investigations of microbial communities. The Gateway provides a platform for microbiology experiments both inside, to improve understanding of interactions between microbes and human habitats, and outside, to improve understanding of microbe-hardware interactions exposed to the space environment.

  1. Evolutionary process of deep-sea bathymodiolus mussels.

    Science.gov (United States)

    Miyazaki, Jun-Ichi; de Oliveira Martins, Leonardo; Fujita, Yuko; Matsumoto, Hiroto; Fujiwara, Yoshihiro

    2010-04-27

    Since the discovery of deep-sea chemosynthesis-based communities, much work has been done to clarify their organismal and environmental aspects. However, major topics remain to be resolved, including when and how organisms invade and adapt to deep-sea environments; whether strategies for invasion and adaptation are shared by different taxa or unique to each taxon; how organisms extend their distribution and diversity; and how they become isolated to speciate in continuous waters. Deep-sea mussels are one of the dominant organisms in chemosynthesis-based communities, thus investigations of their origin and evolution contribute to resolving questions about life in those communities. We investigated worldwide phylogenetic relationships of deep-sea Bathymodiolus mussels and their mytilid relatives by analyzing nucleotide sequences of the mitochondrial cytochrome c oxidase subunit I (COI) and NADH dehydrogenase subunit 4 (ND4) genes. Phylogenetic analysis of the concatenated sequence data showed that mussels of the subfamily Bathymodiolinae from vents and seeps were divided into four groups, and that mussels of the subfamily Modiolinae from sunken wood and whale carcasses assumed the outgroup position and shallow-water modioline mussels were positioned more distantly to the bathymodioline mussels. We provisionally hypothesized the evolutionary history of Bathymodilolus mussels by estimating evolutionary time under a relaxed molecular clock model. Diversification of bathymodioline mussels was initiated in the early Miocene, and subsequently diversification of the groups occurred in the early to middle Miocene. The phylogenetic relationships support the "Evolutionary stepping stone hypothesis," in which mytilid ancestors exploited sunken wood and whale carcasses in their progressive adaptation to deep-sea environments. This hypothesis is also supported by the evolutionary transition of symbiosis in that nutritional adaptation to the deep sea proceeded from extracellular

  2. Evolutionary process of deep-sea bathymodiolus mussels.

    Directory of Open Access Journals (Sweden)

    Jun-Ichi Miyazaki

    Full Text Available BACKGROUND: Since the discovery of deep-sea chemosynthesis-based communities, much work has been done to clarify their organismal and environmental aspects. However, major topics remain to be resolved, including when and how organisms invade and adapt to deep-sea environments; whether strategies for invasion and adaptation are shared by different taxa or unique to each taxon; how organisms extend their distribution and diversity; and how they become isolated to speciate in continuous waters. Deep-sea mussels are one of the dominant organisms in chemosynthesis-based communities, thus investigations of their origin and evolution contribute to resolving questions about life in those communities. METHODOLOGY/PRINCIPAL FINDING: We investigated worldwide phylogenetic relationships of deep-sea Bathymodiolus mussels and their mytilid relatives by analyzing nucleotide sequences of the mitochondrial cytochrome c oxidase subunit I (COI and NADH dehydrogenase subunit 4 (ND4 genes. Phylogenetic analysis of the concatenated sequence data showed that mussels of the subfamily Bathymodiolinae from vents and seeps were divided into four groups, and that mussels of the subfamily Modiolinae from sunken wood and whale carcasses assumed the outgroup position and shallow-water modioline mussels were positioned more distantly to the bathymodioline mussels. We provisionally hypothesized the evolutionary history of Bathymodilolus mussels by estimating evolutionary time under a relaxed molecular clock model. Diversification of bathymodioline mussels was initiated in the early Miocene, and subsequently diversification of the groups occurred in the early to middle Miocene. CONCLUSIONS/SIGNIFICANCE: The phylogenetic relationships support the "Evolutionary stepping stone hypothesis," in which mytilid ancestors exploited sunken wood and whale carcasses in their progressive adaptation to deep-sea environments. This hypothesis is also supported by the evolutionary transition of

  3. Deep imitation learning for 3D navigation tasks.

    Science.gov (United States)

    Hussein, Ahmed; Elyan, Eyad; Gaber, Mohamed Medhat; Jayne, Chrisina

    2018-01-01

    Deep learning techniques have shown success in learning from raw high-dimensional data in various applications. While deep reinforcement learning is recently gaining popularity as a method to train intelligent agents, utilizing deep learning in imitation learning has been scarcely explored. Imitation learning can be an efficient method to teach intelligent agents by providing a set of demonstrations to learn from. However, generalizing to situations that are not represented in the demonstrations can be challenging, especially in 3D environments. In this paper, we propose a deep imitation learning method to learn navigation tasks from demonstrations in a 3D environment. The supervised policy is refined using active learning in order to generalize to unseen situations. This approach is compared to two popular deep reinforcement learning techniques: deep-Q-networks and Asynchronous actor-critic (A3C). The proposed method as well as the reinforcement learning methods employ deep convolutional neural networks and learn directly from raw visual input. Methods for combining learning from demonstrations and experience are also investigated. This combination aims to join the generalization ability of learning by experience with the efficiency of learning by imitation. The proposed methods are evaluated on 4 navigation tasks in a 3D simulated environment. Navigation tasks are a typical problem that is relevant to many real applications. They pose the challenge of requiring demonstrations of long trajectories to reach the target and only providing delayed rewards (usually terminal) to the agent. The experiments show that the proposed method can successfully learn navigation tasks from raw visual input while learning from experience methods fail to learn an effective policy. Moreover, it is shown that active learning can significantly improve the performance of the initially learned policy using a small number of active samples.

  4. Myopes show increased susceptibility to nearwork aftereffects.

    Science.gov (United States)

    Ciuffreda, K J; Wallis, D M

    1998-09-01

    Some aspects of accommodation may be slightly abnormal (or different) in myopes, compared with accommodation in emmetropes and hyperopes. For example, the initial magnitude of accommodative adaptation in the dark after nearwork is greatest in myopes. However, the critical test is to assess this initial accommodative aftereffect and its subsequent decay in the light under more natural viewing conditions with blur-related visual feedback present, if a possible link between this phenomenon and clinical myopia is to be considered. Subjects consisted of adult late- (n = 11) and early-onset (n = 13) myopes, emmetropes (n = 11), and hyperopes (n = 9). The distance-refractive state was assessed objectively using an autorefractor immediately before and after a 10-minute binocular near task at 20 cm (5 diopters [D]). Group results showed that myopes were most susceptible to the nearwork aftereffect. It averaged 0.35 D in initial magnitude, with considerably faster posttask decay to baseline in the early-onset (35 seconds) versus late-onset (63 seconds) myopes. There was no myopic aftereffect in the remaining two refractive groups. The myopes showed particularly striking accommodatively related nearwork aftereffect susceptibility. As has been speculated and found by many others, transient pseudomyopia may cause or be a precursor to permanent myopia or myopic progression. Time-integrated increased retinal defocus causing axial elongation is proposed as a possible mechanism.

  5. Boosted jet identification using particle candidates and deep neural networks

    CERN Document Server

    CMS Collaboration

    2017-01-01

    This note presents developments for the identification of hadronically decaying top quarks using deep neural networks in CMS. A new method that utilizes one dimensional convolutional neural networks based on jet constituent particles is proposed. Alternative methods using boosted decision trees based on jet observables are compared. The new method shows significant improvement in performance.

  6. Deep Learning based Super-Resolution for Improved Action Recognition

    DEFF Research Database (Denmark)

    Nasrollahi, Kamal; Guerrero, Sergio Escalera; Rasti, Pejman

    2015-01-01

    with results of a state-of- the-art deep learning-based super-resolution algorithm, through an alpha-blending approach. The experimental results obtained on down-sampled version of a large subset of Hoolywood2 benchmark database show the importance of the proposed system in increasing the recognition rate...

  7. Real time wave measurements and wave hindcasting in deep waters

    Digital Repository Service at National Institute of Oceanography (India)

    Anand, N.M.; Mandal, S.; SanilKumar, V.; Nayak, B.U.

    Deep water waves off Karwar (lat. 14~'45.1'N, long. 73~'34.8'E) at 75 m water depth pertaining to peak monsoon period have been measured using a Datawell waverider buoy. Measured wave data show that the significant wave height (Hs) predominantly...

  8. Vision in the deep sea.

    Science.gov (United States)

    Warrant, Eric J; Locket, N Adam

    2004-08-01

    The deep sea is the largest habitat on earth. Its three great faunal environments--the twilight mesopelagic zone, the dark bathypelagic zone and the vast flat expanses of the benthic habitat--are home to a rich fauna of vertebrates and invertebrates. In the mesopelagic zone (150-1000 m), the down-welling daylight creates an extended scene that becomes increasingly dimmer and bluer with depth. The available daylight also originates increasingly from vertically above, and bioluminescent point-source flashes, well contrasted against the dim background daylight, become increasingly visible. In the bathypelagic zone below 1000 m no daylight remains, and the scene becomes entirely dominated by point-like bioluminescence. This changing nature of visual scenes with depth--from extended source to point source--has had a profound effect on the designs of deep-sea eyes, both optically and neurally, a fact that until recently was not fully appreciated. Recent measurements of the sensitivity and spatial resolution of deep-sea eyes--particularly from the camera eyes of fishes and cephalopods and the compound eyes of crustaceans--reveal that ocular designs are well matched to the nature of the visual scene at any given depth. This match between eye design and visual scene is the subject of this review. The greatest variation in eye design is found in the mesopelagic zone, where dim down-welling daylight and bio-luminescent point sources may be visible simultaneously. Some mesopelagic eyes rely on spatial and temporal summation to increase sensitivity to a dim extended scene, while others sacrifice this sensitivity to localise pinpoints of bright bioluminescence. Yet other eyes have retinal regions separately specialised for each type of light. In the bathypelagic zone, eyes generally get smaller and therefore less sensitive to point sources with increasing depth. In fishes, this insensitivity, combined with surprisingly high spatial resolution, is very well adapted to the

  9. The deep Canary poleward undercurrent

    Science.gov (United States)

    Velez-Belchi, P. J.; Hernandez-Guerra, A.; González-Pola, C.; Fraile, E.; Collins, C. A.; Machín, F.

    2012-12-01

    Poleward undercurrents are well known features in Eastern Boundary systems. In the California upwelling system (CalCEBS), the deep poleward flow has been observed along the entire outer continental shelf and upper-slope, using indirect methods based on geostrophic estimates and also using direct current measurements. The importance of the poleward undercurrents in the CalCEBS, among others, is to maintain its high productivity by means of the transport of equatorial Pacific waters all the way northward to Vancouver Island and the subpolar gyre but there is also concern about the low oxygen concentration of these waters. However, in the case of the Canary Current Eastern Boundary upwelling system (CanCEBS), there are very few observations of the poleward undercurrent. Most of these observations are short-term mooring records, or drifter trajectories of the upper-slope flow. Hence, the importance of the subsurface poleward flow in the CanCEBS has been only hypothesized. Moreover, due to the large differences between the shape of the coastline and topography between the California and the Canary Current system, the results obtained for the CalCEBS are not completely applicable to the CanCEBS. In this study we report the first direct observations of the continuity of the deep poleward flow of the Canary Deep Poleward undercurrent (CdPU) in the North-Africa sector of the CanCEBS, and one of the few direct observations in the North-Africa sector of the Canary Current eastern boundary. The results indicate that the Canary Island archipelago disrupts the deep poleward undercurrent even at depths where the flow is not blocked by the bathymetry. The deep poleward undercurrent flows west around the eastern-most islands and north east of the Conception Bank to rejoin the intermittent branch that follows the African slope in the Lanzarote Passage. This hypothesis is consistent with the AAIW found west of Lanzarote, as far as 17 W. But also, this hypothesis would be coherent

  10. Image quality assessment using deep convolutional networks

    Science.gov (United States)

    Li, Yezhou; Ye, Xiang; Li, Yong

    2017-12-01

    This paper proposes a method of accurately assessing image quality without a reference image by using a deep convolutional neural network. Existing training based methods usually utilize a compact set of linear filters for learning features of images captured by different sensors to assess their quality. These methods may not be able to learn the semantic features that are intimately related with the features used in human subject assessment. Observing this drawback, this work proposes training a deep convolutional neural network (CNN) with labelled images for image quality assessment. The ReLU in the CNN allows non-linear transformations for extracting high-level image features, providing a more reliable assessment of image quality than linear filters. To enable the neural network to take images of any arbitrary size as input, the spatial pyramid pooling (SPP) is introduced connecting the top convolutional layer and the fully-connected layer. In addition, the SPP makes the CNN robust to object deformations to a certain extent. The proposed method taking an image as input carries out an end-to-end learning process, and outputs the quality of the image. It is tested on public datasets. Experimental results show that it outperforms existing methods by a large margin and can accurately assess the image quality on images taken by different sensors of varying sizes.

  11. Deep inelastic singlet structure functions and scaling violation

    Energy Technology Data Exchange (ETDEWEB)

    Wen-zhu, Li; Bing-xun, Hu

    1984-02-01

    The flavour singlet structure functions of deep inelastic scattering processes can yield more decisive tests of QCD than the non-singlet. We give analytical expression for flavour singlet structure functions through analysing the lepton-nucleon deep inelastic scattering processes by means of QCD and using Jacobi polynomials. This expression contains 4 to 5 parameters and shows the changes of the singlet structure functions with x and Q/sup 2/ very well. In QCD leading order, the conclusion is in reasonable agreement with experimental data.

  12. An adaptive sampling scheme for deep-penetration calculation

    International Nuclear Information System (INIS)

    Wang, Ruihong; Ji, Zhicheng; Pei, Lucheng

    2013-01-01

    As we know, the deep-penetration problem has been one of the important and difficult problems in shielding calculation with Monte Carlo Method for several decades. In this paper, an adaptive Monte Carlo method under the emission point as a sampling station for shielding calculation is investigated. The numerical results show that the adaptive method may improve the efficiency of the calculation of shielding and might overcome the under-estimation problem easy to happen in deep-penetration calculation in some degree

  13. Deep iCrawl: An Intelligent Vision-Based Deep Web Crawler

    OpenAIRE

    R.Anita; V.Ganga Bharani; N.Nityanandam; Pradeep Kumar Sahoo

    2011-01-01

    The explosive growth of World Wide Web has posed a challenging problem in extracting relevant data. Traditional web crawlers focus only on the surface web while the deep web keeps expanding behind the scene. Deep web pages are created dynamically as a result of queries posed to specific web databases. The structure of the deep web pages makes it impossible for traditional web crawlers to access deep web contents. This paper, Deep iCrawl, gives a novel and vision-based app...

  14. Ancient bacteria show evidence of DNA repair

    DEFF Research Database (Denmark)

    Johnson, Sarah Stewart; Hebsgaard, Martin B; Christensen, Torben R

    2007-01-01

    -term survival of bacteria sealed in frozen conditions for up to one million years. Our results show evidence of bacterial survival in samples up to half a million years in age, making this the oldest independently authenticated DNA to date obtained from viable cells. Additionally, we find strong evidence...... geological timescales. There has been no direct evidence in ancient microbes for the most likely mechanism, active DNA repair, or for the metabolic activity necessary to sustain it. In this paper, we couple PCR and enzymatic treatment of DNA with direct respiration measurements to investigate long...... that this long-term survival is closely tied to cellular metabolic activity and DNA repair that over time proves to be superior to dormancy as a mechanism in sustaining bacteria viability....

  15. Microbiological and environmental issues in show caves.

    Science.gov (United States)

    Saiz-Jimenez, Cesareo

    2012-07-01

    Cultural tourism expanded in the last half of the twentieth century, and the interest of visitors has come to include caves containing archaeological remains. Some show caves attracted mass tourism, and economical interests prevailed over conservation, which led to a deterioration of the subterranean environment and the rock art. The presence and the role of microorganisms in caves is a topic that is often ignored in cave management. Knowledge of the colonisation patterns, the dispersion mechanisms, and the effect on human health and, when present, over rock art paintings of these microorganisms is of the utmost importance. In this review the most recent advances in the study of microorganisms in caves are presented, together with the environmental implications of the findings.

  16. Exploring orange peel treatment with deep eutectic solvents and diluted organic acids

    NARCIS (Netherlands)

    van den Bruinhorst, A.; Kouris, P.; Timmer, J.M.K.; de Croon, M.H.J.M.; Kroon, M.C.

    2016-01-01

    The disintegration of orange peel waste in deep eutectic solvents and diluted organic acids is presented in this work. The albedo and flavedo layers of the peel were studied separately, showing faster disintegration of the latter. Addition of water to the deep eutectic solvents lowered the amount of

  17. Giant lipoma arising from deep lobe of the parotid gland

    Directory of Open Access Journals (Sweden)

    Hsu Ying-Che

    2006-06-01

    Full Text Available Abstract Background Lipomas are common benign soft tissue neoplasms but they are found very rarely in the deep lobe of parotid gland. Surgical intervention in these tumors is challenging because of the proximity of the facial nerve, and thus knowledge of the anatomy and meticulous surgical technique are essential. Case presentation A 71-year-old female presented with a large asymptomatic mass, which had occupied the left facial area for over the past fifteen years, and she requested surgical excision for a cosmetically better facial appearance. The computed tomography (CT scan showed a well-defined giant lipoma arising from the left deep parotid gland. The lipoma was successfully enucleated after full exposure and mobilization of the overlying facial nerve branches. The surgical specimen measured 9 × 6 cm in size, and histopathology revealed fibrolipoma. The patient experienced an uneventful recovery, with a satisfying facial contour and intact facial nerve function. Conclusion Giant lipomas involving the deep parotid lobe are extremely rare. The high-resolution CT scan provides an accurate and cost-effective preoperative investigative method. Surgical management of deep lobe lipoma should be performed by experienced surgeons due to the need for meticulous dissection of the facial nerve branches. Superficial parotidectomy before deep lobe lipoma removal may be unnecessary in selected cases because preservation of the superficial lobe may contribute to a better aesthetic and functional result.

  18. Benchmarking Deep Learning Models on Large Healthcare Datasets.

    Science.gov (United States)

    Purushotham, Sanjay; Meng, Chuizheng; Che, Zhengping; Liu, Yan

    2018-06-04

    Deep learning models (aka Deep Neural Networks) have revolutionized many fields including computer vision, natural language processing, speech recognition, and is being increasingly used in clinical healthcare applications. However, few works exist which have benchmarked the performance of the deep learning models with respect to the state-of-the-art machine learning models and prognostic scoring systems on publicly available healthcare datasets. In this paper, we present the benchmarking results for several clinical prediction tasks such as mortality prediction, length of stay prediction, and ICD-9 code group prediction using Deep Learning models, ensemble of machine learning models (Super Learner algorithm), SAPS II and SOFA scores. We used the Medical Information Mart for Intensive Care III (MIMIC-III) (v1.4) publicly available dataset, which includes all patients admitted to an ICU at the Beth Israel Deaconess Medical Center from 2001 to 2012, for the benchmarking tasks. Our results show that deep learning models consistently outperform all the other approaches especially when the 'raw' clinical time series data is used as input features to the models. Copyright © 2018 Elsevier Inc. All rights reserved.

  19. Quantification of deep medullary veins at 7 T brain MRI

    Energy Technology Data Exchange (ETDEWEB)

    Kuijf, Hugo J.; Viergever, Max A.; Vincken, Koen L. [University Medical Center Utrecht, Image Sciences Institute, Utrecht (Netherlands); Bouvy, Willem H.; Razoux Schultz, Tom B.; Biessels, Geert Jan [University Medical Center Utrecht, Department of Neurology, Brain Center Rudolf Magnus, Utrecht (Netherlands); Zwanenburg, Jaco J.M. [University Medical Center Utrecht, Image Sciences Institute, Utrecht (Netherlands); University Medical Center Utrecht, Department of Radiology, Utrecht (Netherlands)

    2016-10-15

    Deep medullary veins support the venous drainage of the brain and may display abnormalities in the context of different cerebrovascular diseases. We present and evaluate a method to automatically detect and quantify deep medullary veins at 7 T. Five participants were scanned twice, to assess the robustness and reproducibility of manual and automated vein detection. Additionally, the method was evaluated on 24 participants to demonstrate its application. Deep medullary veins were assessed within an automatically created region-of-interest around the lateral ventricles, defined such that all veins must intersect it. A combination of vesselness, tubular tracking, and hysteresis thresholding located individual veins, which were quantified by counting and computing (3-D) density maps. Visual assessment was time-consuming (2 h/scan), with an intra-/inter-observer agreement on absolute vein count of ICC = 0.76 and 0.60, respectively. The automated vein detection showed excellent inter-scan reproducibility before (ICC = 0.79) and after (ICC = 0.88) visually censoring false positives. It had a positive predictive value of 71.6 %. Imaging at 7 T allows visualization and quantification of deep medullary veins. The presented method offers fast and reliable automated assessment of deep medullary veins. (orig.)

  20. Shakeout: A New Approach to Regularized Deep Neural Network Training.

    Science.gov (United States)

    Kang, Guoliang; Li, Jun; Tao, Dacheng

    2018-05-01

    Recent years have witnessed the success of deep neural networks in dealing with a plenty of practical problems. Dropout has played an essential role in many successful deep neural networks, by inducing regularization in the model training. In this paper, we present a new regularized training approach: Shakeout. Instead of randomly discarding units as Dropout does at the training stage, Shakeout randomly chooses to enhance or reverse each unit's contribution to the next layer. This minor modification of Dropout has the statistical trait: the regularizer induced by Shakeout adaptively combines , and regularization terms. Our classification experiments with representative deep architectures on image datasets MNIST, CIFAR-10 and ImageNet show that Shakeout deals with over-fitting effectively and outperforms Dropout. We empirically demonstrate that Shakeout leads to sparser weights under both unsupervised and supervised settings. Shakeout also leads to the grouping effect of the input units in a layer. Considering the weights in reflecting the importance of connections, Shakeout is superior to Dropout, which is valuable for the deep model compression. Moreover, we demonstrate that Shakeout can effectively reduce the instability of the training process of the deep architecture.

  1. Deep Corals, Deep Learning: Moving the Deep Net Towards Real-Time Image Annotation

    OpenAIRE

    Lea-Anne Henry; Sankha S. Mukherjee; Neil M. Roberston; Laurence De Clippele; J. Murray Roberts

    2016-01-01

    The mismatch between human capacity and the acquisition of Big Data such as Earth imagery undermines commitments to Convention on Biological Diversity (CBD) and Aichi targets. Artificial intelligence (AI) solutions to Big Data issues are urgently needed as these could prove to be faster, more accurate, and cheaper. Reducing costs of managing protected areas in remote deep waters and in the High Seas is of great importance, and this is a realm where autonomous technology will be transformative.

  2. Invited talk: Deep Learning Meets Physics

    CERN Multimedia

    CERN. Geneva

    2018-01-01

    Deep Learning has emerged as one of the most successful fields of machine learning and artificial intelligence with overwhelming success in industrial speech, text and vision benchmarks. Consequently it evolved into the central field of research for IT giants like Google, facebook, Microsoft, Baidu, and Amazon. Deep Learning is founded on novel neural network techniques, the recent availability of very fast computers, and massive data sets. In its core, Deep Learning discovers multiple levels of abstract representations of the input. The main obstacle to learning deep neural networks is the vanishing gradient problem. The vanishing gradient impedes credit assignment to the first layers of a deep network or to early elements of a sequence, therefore limits model selection. Major advances in Deep Learning can be related to avoiding the vanishing gradient like stacking, ReLUs, residual networks, highway networks, and LSTM. For Deep Learning, we suggested self-normalizing neural networks (SNNs) which automatica...

  3. Learning Multimodal Deep Representations for Crowd Anomaly Event Detection

    Directory of Open Access Journals (Sweden)

    Shaonian Huang

    2018-01-01

    Full Text Available Anomaly event detection in crowd scenes is extremely important; however, the majority of existing studies merely use hand-crafted features to detect anomalies. In this study, a novel unsupervised deep learning framework is proposed to detect anomaly events in crowded scenes. Specifically, low-level visual features, energy features, and motion map features are simultaneously extracted based on spatiotemporal energy measurements. Three convolutional restricted Boltzmann machines are trained to model the mid-level feature representation of normal patterns. Then a multimodal fusion scheme is utilized to learn the deep representation of crowd patterns. Based on the learned deep representation, a one-class support vector machine model is used to detect anomaly events. The proposed method is evaluated using two available public datasets and compared with state-of-the-art methods. The experimental results show its competitive performance for anomaly event detection in video surveillance.

  4. Plastic microfibre ingestion by deep-sea organisms

    Science.gov (United States)

    Taylor, M. L.; Gwinnett, C.; Robinson, L. F.; Woodall, L. C.

    2016-09-01

    Plastic waste is a distinctive indicator of the world-wide impact of anthropogenic activities. Both macro- and micro-plastics are found in the ocean, but as yet little is known about their ultimate fate and their impact on marine ecosystems. In this study we present the first evidence that microplastics are already becoming integrated into deep-water organisms. By examining organisms that live on the deep-sea floor we show that plastic microfibres are ingested and internalised by members of at least three major phyla with different feeding mechanisms. These results demonstrate that, despite its remote location, the deep sea and its fragile habitats are already being exposed to human waste to the extent that diverse organisms are ingesting microplastics.

  5. [Deep continuous palliative sedation in the Opinion adopted by the Italian National Bioethics Committee (Deep palliative sedation)].

    Science.gov (United States)

    Cembrani, Fabio

    2016-01-01

    The Author examines the recent opinion delivered by the Italian National Committee for Bioethics on deep palliative sedation. In particular, it examines its strengths and ample shade that show its ideology, once again, in contrast with the right of every human being to die with dignity.

  6. NASA GIBS Use in Live Planetarium Shows

    Science.gov (United States)

    Emmart, C. B.

    2015-12-01

    The American Museum of Natural History's Hayden Planetarium was rebuilt in year 2000 as an immersive theater for scientific data visualization to show the universe in context to our planet. Specific astrophysical movie productions provide the main daily programming, but interactive control software, developed at AMNH allows immersive presentation within a data aggregation of astronomical catalogs called the Digital Universe 3D Atlas. Since 2006, WMS globe browsing capabilities have been built into a software development collaboration with Sweden's Linkoping University (LiU). The resulting Uniview software, now a product of the company SCISS, is operated by about fifty planetariums around that world with ability to network amongst the sites for global presentations. Public presentation of NASA GIBS has allowed authoritative narratives to be presented within the range of data available in context to other sources such as Science on a Sphere, NASA Earth Observatory and Google Earth KML resources. Specifically, the NOAA supported World Views Network conducted a series of presentations across the US that focused on local ecological issues that could then be expanded in the course of presentation to national and global scales of examination. NASA support of for GIBS resources in an easy access multi scale streaming format like WMS has tremendously enabled particularly facile presentations of global monitoring like never before. Global networking of theaters for distributed presentations broadens out the potential for impact of this medium. Archiving and refinement of these presentations has already begun to inform new types of documentary productions that examine pertinent, global interdependency topics.

  7. Geoscience is Important? Show Me Why

    Science.gov (United States)

    Boland, M. A.

    2017-12-01

    "The public" is not homogenous and no single message or form of messaging will connect the entire public with the geosciences. One approach to promoting trust in, and engagement with, the geosciences is to identify specific sectors of the public and then develop interactions and communication products that are immediately relevant to that sector's interests. If the content and delivery are appropriate, this approach empowers people to connect with the geosciences on their own terms and to understand the relevance of the geosciences to their own situation. Federal policy makers are a distinct and influential subgroup of the general public. In preparation for the 2016 presidential election, the American Geosciences Institute (AGI) in collaboration with its 51 member societies prepared Geoscience for America's Critical Needs: Invitation to a National Dialogue, a document that identified major geoscience policy issues that should be addressed in a national policy platform. Following the election, AGI worked with eight other geoscience societies to develop Geoscience Policy Recommendations for the New Administration and the 115th Congress, which outlines specific policy actions to address national issues. State and local decision makers are another important subgroup of the public. AGI has developed online content, factsheets, and case studies with different levels of technical complexity so people can explore societally-relevant geoscience topics at their level of technical proficiency. A related webinar series is attracting a growing worldwide audience from many employment sectors. Partnering with government agencies and other scientific and professional societies has increased the visibility and credibility of these information products with our target audience. Surveys and other feedback show that these products are raising awareness of the geosciences and helping to build reciprocal relationships between geoscientists and decision makers. The core message of all

  8. Bacteriophages show promise as antimicrobial agents.

    Science.gov (United States)

    Alisky, J; Iczkowski, K; Rapoport, A; Troitsky, N

    1998-01-01

    The emergence of antibiotic-resistant bacteria has prompted interest in alternatives to conventional drugs. One possible option is to use bacteriophages (phage) as antimicrobial agents. We have conducted a literature review of all Medline citations from 1966-1996 that dealt with the therapeutic use of phage. There were 27 papers from Poland, the Soviet Union, Britain and the U.S.A. The Polish and Soviets administered phage orally, topically or systemically to treat a wide variety of antibiotic-resistant pathogens in both adults and children. Infections included suppurative wound infections, gastroenteritis, sepsis, osteomyelitis, dermatitis, empyemas and pneumonia; pathogens included Staphylococcus, Streptococcus, Klebsiella, Escherichia, Proteus, Pseudomonas, Shigella and Salmonella spp. Overall, the Polish and Soviets reported success rates of 80-95% for phage therapy, with rare, reversible gastrointestinal or allergic side effects. However, efficacy of phage was determined almost exclusively by qualitative clinical assessment of patients, and details of dosages and clinical criteria were very sketchy. There were also six British reports describing controlled trials of phage in animal models (mice, guinea pigs and livestock), measuring survival rates and other objective criteria. All of the British studies raised phage against specific pathogens then used to create experimental infections. Demonstrable efficacy against Escherichia, Acinetobacter, Pseudomonas and Staphylococcus spp. was noted in these model systems. Two U.S. papers dealt with improving the bioavailability of phage. Phage is sequestered in the spleen and removed from circulation. This can be overcome by serial passage of phage through mice to isolate mutants that resist sequestration. In conclusion, bacteriophages may show promise for treating antibiotic resistant pathogens. To facilitate further progress, directions for future research are discussed and a directory of authors from the reviewed

  9. Deep Learning with Dynamic Spiking Neurons and Fixed Feedback Weights.

    Science.gov (United States)

    Samadi, Arash; Lillicrap, Timothy P; Tweed, Douglas B

    2017-03-01

    Recent work in computer science has shown the power of deep learning driven by the backpropagation algorithm in networks of artificial neurons. But real neurons in the brain are different from most of these artificial ones in at least three crucial ways: they emit spikes rather than graded outputs, their inputs and outputs are related dynamically rather than by piecewise-smooth functions, and they have no known way to coordinate arrays of synapses in separate forward and feedback pathways so that they change simultaneously and identically, as they do in backpropagation. Given these differences, it is unlikely that current deep learning algorithms can operate in the brain, but we that show these problems can be solved by two simple devices: learning rules can approximate dynamic input-output relations with piecewise-smooth functions, and a variation on the feedback alignment algorithm can train deep networks without having to coordinate forward and feedback synapses. Our results also show that deep spiking networks learn much better if each neuron computes an intracellular teaching signal that reflects that cell's nonlinearity. With this mechanism, networks of spiking neurons show useful learning in synapses at least nine layers upstream from the output cells and perform well compared to other spiking networks in the literature on the MNIST digit recognition task.

  10. Exploring frontiers of the deep biosphere through scientific ocean drilling

    Science.gov (United States)

    Inagaki, F.; D'Hondt, S.; Hinrichs, K. U.

    2015-12-01

    Since the first deep biosphere-dedicated Ocean Drilling Program (ODP) Leg 201 using the US drill ship JOIDES Resolution in 2002, scientific ocean drilling has offered unique opportunities to expand our knowledge of the nature and extent of the deep biosphere. The latest estimate of the global subseafloor microbial biomass is ~1029cells, accounting for 4 Gt of carbon and ~1% of the Earth's total living biomass. The subseafloor microbial communities are evolutionarily diverse and their metabolic rates are extraordinarily slow. Nevertheless, accumulating activity most likely plays a significant role in elemental cycles over geological time. In 2010, during Integrated Ocean Drilling Program (IODP) Expedition 329, the JOIDES Resolutionexplored the deep biosphere in the open-ocean South Pacific Gyre—the largest oligotrophic province on our planet. During Expedition 329, relatively high concentrations of dissolved oxygen and significantly low biomass of microbial populations were observed in the entire sediment column, indicating that (i) there is no limit to life in open-ocean sediment and (ii) a significant amount of oxygen reaches through the sediment to the upper oceanic crust. This "deep aerobic biosphere" inhabits the sediment throughout up to ~37 percent of the world's oceans. The remaining ~63 percent of the oceans is comprised of higher productivity areas that contain the "deep anaerobic biosphere". In 2012, during IODP Expedition 337, the Japanese drill ship Chikyu explored coal-bearing sediments down to 2,466 meters below the seafloor off the Shimokita Peninsula, Japan. Geochemical and microbiological analyses consistently showed the occurrence of methane-producing communities associated with the coal beds. Cell concentrations in deep sediments were notably lower than those expected from the global regression line, implying that the bottom of the deep biosphere is approached in these beds. Taxonomic composition of the deep coal-bearing communities profoundly

  11. Deep remission: a new concept?

    Science.gov (United States)

    Colombel, Jean-Frédéric; Louis, Edouard; Peyrin-Biroulet, Laurent; Sandborn, William J; Panaccione, Remo

    2012-01-01

    Crohn's disease (CD) is a chronic inflammatory disorder characterized by periods of clinical remission alternating with periods of relapse defined by recurrent clinical symptoms. Persistent inflammation is believed to lead to progressive bowel damage over time, which manifests with the development of strictures, fistulae and abscesses. These disease complications frequently lead to a need for surgical resection, which in turn leads to disability. So CD can be characterized as a chronic, progressive, destructive and disabling disease. In rheumatoid arthritis, treatment paradigms have evolved beyond partial symptom control alone toward the induction and maintenance of sustained biological remission, also known as a 'treat to target' strategy, with the goal of improving long-term disease outcomes. In CD, there is currently no accepted, well-defined, comprehensive treatment goal that entails the treatment of both clinical symptoms and biologic inflammation. It is important that such a treatment concept begins to evolve for CD. A treatment strategy that delays or halts the progression of CD to increasing damage and disability is a priority. As a starting point, a working definition of sustained deep remission (that includes long-term biological remission and symptom control) with defined patient outcomes (including no disease progression) has been proposed. The concept of sustained deep remission represents a goal for CD management that may still evolve. It is not clear if the concept also applies to ulcerative colitis. Clinical trials are needed to evaluate whether treatment algorithms that tailor therapy to achieve deep remission in patients with CD can prevent disease progression and disability. Copyright © 2012 S. Karger AG, Basel.

  12. Topics in deep inelastic scattering

    International Nuclear Information System (INIS)

    Wandzura, S.M.

    1977-01-01

    Several topics in deep inelastic lepton--nucleon scattering are discussed, with emphasis on the structure functions appearing in polarized experiments. The major results are: infinite set of new sum rules reducing the number of independent spin dependent structure functions (for electroproduction) from two to one; the application of the techniques of Nachtmann to extract the coefficients appearing in the Wilson operator product expansion; and radiative corrections to the Wilson coefficients of free field theory. Also discussed are the use of dimensional regularization to simplify the calculation of these radiative corrections

  13. An OSSE Study for Deep Argo Array using the GFDL Ensemble Coupled Data Assimilation System

    Science.gov (United States)

    Chang, You-Soon; Zhang, Shaoqing; Rosati, Anthony; Vecchi, Gabriel A.; Yang, Xiaosong

    2018-03-01

    An observing system simulation experiment (OSSE) using an ensemble coupled data assimilation system was designed to investigate the impact of deep ocean Argo profile assimilation in a biased numerical climate system. Based on the modern Argo observational array and an artificial extension to full depth, "observations" drawn from one coupled general circulation model (CM2.0) were assimilated into another model (CM2.1). Our results showed that coupled data assimilation with simultaneous atmospheric and oceanic constraints plays a significant role in preventing deep ocean drift. However, the extension of the Argo array to full depth did not significantly improve the quality of the oceanic climate estimation within the bias magnitude in the twin experiment. Even in the "identical" twin experiment for the deep Argo array from the same model (CM2.1) with the assimilation model, no significant changes were shown in the deep ocean, such as in the Atlantic meridional overturning circulation and the Antarctic bottom water cell. The small ensemble spread and corresponding weak constraints by the deep Argo profiles with medium spatial and temporal resolution may explain why the deep Argo profiles did not improve the deep ocean features in the assimilation system. Additional studies using different assimilation methods with improved spatial and temporal resolution of the deep Argo array are necessary in order to more thoroughly understand the impact of the deep Argo array on the assimilation system.

  14. Classification of Exacerbation Frequency in the COPDGene Cohort Using Deep Learning with Deep Belief Networks.

    Science.gov (United States)

    Ying, Jun; Dutta, Joyita; Guo, Ning; Hu, Chenhui; Zhou, Dan; Sitek, Arkadiusz; Li, Quanzheng

    2016-12-21

    This study aims to develop an automatic classifier based on deep learning for exacerbation frequency in patients with chronic obstructive pulmonary disease (COPD). A threelayer deep belief network (DBN) with two hidden layers and one visible layer was employed to develop classification models and the models' robustness to exacerbation was analyzed. Subjects from the COPDGene cohort were labeled with exacerbation frequency, defined as the number of exacerbation events per year. 10,300 subjects with 361 features each were included in the analysis. After feature selection and parameter optimization, the proposed classification method achieved an accuracy of 91.99%, using a 10-fold cross validation experiment. The analysis of DBN weights showed that there was a good visual spatial relationship between the underlying critical features of different layers. Our findings show that the most sensitive features obtained from the DBN weights are consistent with the consensus showed by clinical rules and standards for COPD diagnostics. We thus demonstrate that DBN is a competitive tool for exacerbation risk assessment for patients suffering from COPD.

  15. On the dragnosis of deep vein thrombosis

    International Nuclear Information System (INIS)

    Olsson, C.-G.

    1979-01-01

    Clinical and laboratory diagnostic methods were studied in 301 consecutive patients with suspected deep vein thrombosis (DVT). Unexpectedly, phlebography (the reference method) was found to cause DVT in estimated 48 % of patients without initial DVT. Using a new type of contrast medium, however, no thrombotic complications were found. - Neither clinical examination nor plethysmography were found to give reliable results. Using a modified technique for radioisotope detection, high sensitivity to DVT was found with the 125 I-fibrinogen uptake test (within 2 days) and a newly developed 99 Tcsup(m)-plasmin test (within one hour). Since both tests showed low specificity, they are reliable as screening tests to exclude DVT, but not as independent diagnostic methods. (author)

  16. Deep venous thrombophlebitis following aortoiliac reconstructive surgery

    International Nuclear Information System (INIS)

    Reilly, M.K.; McCabe, C.J.; Abbott, W.M.; Brewster, D.C.; Moncure, A.C.; Reidy, N.C.; Darling, R.C.

    1982-01-01

    One hundred patients undergoing elective aortic surgery were scanned prospectively for development of deep venous thrombosis (DVT). The incidence of DVT in this population was 13%. Eleven patients showed only calf vein thrombosis on venography, whereas two had occlusive iliofemoral thrombus. The correlation between venous Doppler ultrasound and venography was 80%. More importantly, Doppler examination correctly identified both patients with occlusive thrombus. Fibrinogen scanning was associated with a false-positive rate of 31%. Only one patient suffered a nonfatal pulmonary embolus. Fibrinogen scanning has an unacceptably high false-positive rate; however, Doppler ultrasound will identify significant occlusive thrombus without a high false-positive rate. The low incidence of pulmonary emboli does not warrant such definitive measures as prophylactic vena caval interruption

  17. Deep hierarchical attention network for video description

    Science.gov (United States)

    Li, Shuohao; Tang, Min; Zhang, Jun

    2018-03-01

    Pairing video to natural language description remains a challenge in computer vision and machine translation. Inspired by image description, which uses an encoder-decoder model for reducing visual scene into a single sentence, we propose a deep hierarchical attention network for video description. The proposed model uses convolutional neural network (CNN) and bidirectional LSTM network as encoders while a hierarchical attention network is used as the decoder. Compared to encoder-decoder models used in video description, the bidirectional LSTM network can capture the temporal structure among video frames. Moreover, the hierarchical attention network has an advantage over single-layer attention network on global context modeling. To make a fair comparison with other methods, we evaluate the proposed architecture with different types of CNN structures and decoders. Experimental results on the standard datasets show that our model has a more superior performance than the state-of-the-art techniques.

  18. Optimizing Chemical Reactions with Deep Reinforcement Learning.

    Science.gov (United States)

    Zhou, Zhenpeng; Li, Xiaocheng; Zare, Richard N

    2017-12-27

    Deep reinforcement learning was employed to optimize chemical reactions. Our model iteratively records the results of a chemical reaction and chooses new experimental conditions to improve the reaction outcome. This model outperformed a state-of-the-art blackbox optimization algorithm by using 71% fewer steps on both simulations and real reactions. Furthermore, we introduced an efficient exploration strategy by drawing the reaction conditions from certain probability distributions, which resulted in an improvement on regret from 0.062 to 0.039 compared with a deterministic policy. Combining the efficient exploration policy with accelerated microdroplet reactions, optimal reaction conditions were determined in 30 min for the four reactions considered, and a better understanding of the factors that control microdroplet reactions was reached. Moreover, our model showed a better performance after training on reactions with similar or even dissimilar underlying mechanisms, which demonstrates its learning ability.

  19. GMSK Modulation for Deep Space Applications

    Science.gov (United States)

    Shambayati, Shervin; Lee, Dennis K.

    2012-01-01

    Due to scarcity of spectrum at 8.42 GHz deep space Xband allocation, many deep space missions are now considering the use of higher order modulation schemes instead of the traditional binary phase shift keying (BPSK). One such scheme is pre-coded Gaussian minimum shift keying (GMSK). GMSK is an excellent candidate for deep space missions. GMSK is a constant envelope, bandwidth efficien modulation whose frame error rate (FER) performance with perfect carrier tracking and proper receiver structure is nearly identical to that of BPSK. There are several issues that need to be addressed with GMSK however. Specificall, we are interested in the combined effects of spectrum limitations and receiver structure on the coded performance of the X-band link using GMSK. The receivers that are typically used for GMSK demodulations are variations on offset quadrature phase shift keying (OQPSK) receivers. In this paper we consider three receivers: the standard DSN OQPSK receiver, DSN OQPSK receiver with filte ed input, and an optimum OQPSK receiver with filte ed input. For the DSN OQPSK receiver we show experimental results with (8920, 1/2), (8920, 1/3) and (8920, 1/6) turbo codes in terms of their error rate performance. We also consider the tracking performance of this receiver as a function of data rate, channel code and the carrier loop signal-to-noise ratio (SNR). For the other two receivers we derive theoretical results that will show that for a given loop bandwidth, a receiver structure, and a channel code, there is a lower data rate limit on the GMSK below which a higher SNR than what is required to achieve the required FER on the link is needed. These limits stem from the minimum loop signal-to-noise ratio requirements on the receivers for achieving lock. As a result of this, for a given channel code and a given FER, there could be a gap between the maximum data rate that BPSK can support without violating the spectrum limits and the minimum data rate that GMSK can support

  20. Deep bottleneck features for spoken language identification.

    Directory of Open Access Journals (Sweden)

    Bing Jiang

    Full Text Available A key problem in spoken language identification (LID is to design effective representations which are specific to language information. For example, in recent years, representations based on both phonotactic and acoustic features have proven their effectiveness for LID. Although advances in machine learning have led to significant improvements, LID performance is still lacking, especially for short duration speech utterances. With the hypothesis that language information is weak and represented only latently in speech, and is largely dependent on the statistical properties of the speech content, existing representations may be insufficient. Furthermore they may be susceptible to the variations caused by different speakers, specific content of the speech segments, and background noise. To address this, we propose using Deep Bottleneck Features (DBF for spoken LID, motivated by the success of Deep Neural Networks (DNN in speech recognition. We show that DBFs can form a low-dimensional compact representation of the original inputs with a powerful descriptive and discriminative capability. To evaluate the effectiveness of this, we design two acoustic models, termed DBF-TV and parallel DBF-TV (PDBF-TV, using a DBF based i-vector representation for each speech utterance. Results on NIST language recognition evaluation 2009 (LRE09 show significant improvements over state-of-the-art systems. By fusing the output of phonotactic and acoustic approaches, we achieve an EER of 1.08%, 1.89% and 7.01% for 30 s, 10 s and 3 s test utterances respectively. Furthermore, various DBF configurations have been extensively evaluated, and an optimal system proposed.

  1. Deep groundwater flow at Palmottu

    International Nuclear Information System (INIS)

    Niini, H.; Vesterinen, M.; Tuokko, T.

    1993-01-01

    Further observations, measurements, and calculations aimed at determining the groundwater flow regimes and periodical variations in flow at deeper levels were carried out in the Lake Palmottu (a natural analogue study site for radioactive waste disposal in southwestern Finland) drainage basin. These water movements affect the migration of radionuclides from the Palmottu U-Th deposit. The deep water flow is essentially restricted to the bedrock fractures which developed under, and are still affected by, the stress state of the bedrock. Determination of the detailed variations was based on fracture-tectonic modelling of the 12 most significant underground water-flow channels that cross the surficial water of the Palmottu area. According to the direction of the hydraulic gradient the deep water flow is mostly outwards from the Palmottu catchment but in the westernmost section it is partly towards the centre. Estimation of the water flow through the U-Th deposit by the water-balance method is still only approximate and needs continued observation series and improved field measurements

  2. Deep ocean model penetrator experiments

    International Nuclear Information System (INIS)

    Freeman, T.J.; Burdett, J.R.F.

    1986-01-01

    Preliminary trials of experimental model penetrators in the deep ocean have been conducted as an international collaborative exercise by participating members (national bodies and the CEC) of the Engineering Studies Task Group of the Nuclear Energy Agency's Seabed Working Group. This report describes and gives the results of these experiments, which were conducted at two deep ocean study areas in the Atlantic: Great Meteor East and the Nares Abyssal Plain. Velocity profiles of penetrators of differing dimensions and weights have been determined as they free-fell through the water column and impacted the sediment. These velocity profiles are used to determine the final embedment depth of the penetrators and the resistance to penetration offered by the sediment. The results are compared with predictions of embedment depth derived from elementary models of a penetrator impacting with a sediment. It is tentatively concluded that once the resistance to penetration offered by a sediment at a particular site has been determined, this quantity can be used to sucessfully predict the embedment that penetrators of differing sizes and weights would achieve at the same site

  3. Academic Training: Deep Space Telescopes

    CERN Multimedia

    Françoise Benz

    2006-01-01

    2005-2006 ACADEMIC TRAINING PROGRAMME LECTURE SERIES 20, 21, 22, 23, 24 February from 11:00 to 12:00 - Council Chamber on 20, 21, 23, 24 February, TH Auditorium, bldg 4 - 3-006, on 22 February Deep Space Telescopes G. BIGNAMI / CNRS, Toulouse, F & Univ. di Pavia, I The short series of seminars will address results and aims of current and future space astrophysics as the cultural framework for the development of deep space telescopes. It will then present such new tools, as they are currently available to, or imagined by, the scientific community, in the context of the science plans of ESA and of all major world space agencies. Ground-based astronomy, in the 400 years since Galileo's telescope, has given us a profound phenomenological comprehension of our Universe, but has traditionally been limited to the narrow band(s) to which our terrestrial atmosphere is transparent. Celestial objects, however, do not care about our limitations, and distribute most of the information about their physics thro...

  4. Enhanced deep ocean ventilation and oxygenation with global warming

    Science.gov (United States)

    Froelicher, T. L.; Jaccard, S.; Dunne, J. P.; Paynter, D.; Gruber, N.

    2014-12-01

    Twenty-first century coupled climate model simulations, observations from the recent past, and theoretical arguments suggest a consistent trend towards warmer ocean temperatures and fresher polar surface oceans in response to increased radiative forcing resulting in increased upper ocean stratification and reduced ventilation and oxygenation of the deep ocean. Paleo-proxy records of the warming at the end of the last ice age, however, suggests a different outcome, namely a better ventilated and oxygenated deep ocean with global warming. Here we use a four thousand year global warming simulation from a comprehensive Earth System Model (GFDL ESM2M) to show that this conundrum is a consequence of different rates of warming and that the deep ocean is actually better ventilated and oxygenated in a future warmer equilibrated climate consistent with paleo-proxy records. The enhanced deep ocean ventilation in the Southern Ocean occurs in spite of increased positive surface buoyancy fluxes and a constancy of the Southern Hemisphere westerly winds - circumstances that would otherwise be expected to lead to a reduction in deep ocean ventilation. This ventilation recovery occurs through a global scale interaction of the Atlantic Meridional Overturning Circulation undergoing a multi-centennial recovery after an initial century of transient decrease and transports salinity-rich waters inform the subtropical surface ocean to the Southern Ocean interior on multi-century timescales. The subsequent upwelling of salinity-rich waters in the Southern Ocean strips away the freshwater cap that maintains vertical stability and increases open ocean convection and the formation of Antarctic Bottom Waters. As a result, the global ocean oxygen content and the nutrient supply from the deep ocean to the surface are higher in a warmer ocean. The implications for past and future changes in ocean heat and carbon storage will be discussed.

  5. Waste disposal in the deep ocean: An overview

    International Nuclear Information System (INIS)

    O'Connor, T.P.; Kester, D.R.; Burt, W.V.; Capuzzo, J.M.; Park, P.K.; Duedall, I.W.

    1985-01-01

    Incineration at sea, industrial and sewage waste disposal in the surface mixing zone, and disposal of low-level nuclear wastes, obsolete munitions, and nerve gas onto the seafloor have been the main uses of the deep sea for waste management. In 1981 the wastes disposed of in the deep sea consisted of 48 X 10/sup 4/ t of liquid industrial wastes and 2 X 10/sup 4/ t of sewage sludge by the United States; 1.5 X 10/sup 4/ t (solids) of sewage sludge by the Federal Republic of German; 5300 t of liquid industrial wastes by Denmark; 99 t of solid industrial wastes by the United Kingdom; and 9400 t of low-level radioactive wastes by several European countries. Also in 1981 at-sea incineration of slightly more than 10/sup 5/ t of organic wastes from Belgium, France, the Federal Republic of Germany, the Netherlands, Norway, Sweden, and the United Kingdom was carried out in the North Sea. Unique oceanographic features of the deep sea include its large dilution capacity; the long residence time of deep-sea water (on the order of 10/sup 2/ y); low biological productivity in the surface water of the open ocean (≅50 g m/sup -2/ of carbon per year); the existence of an oxygen minimum zone at several hundred meters deep in the mid-latitudes; and the abyssal-clay regions showing sedimentary records of tens of millions of years of slow, uninterrupted deposition of fine-grained clay. Any deep-sea waste disposal strategy must take into account oceanic processes and current scientific knowledge in order to attain a safe solution that will last for centuries

  6. Deep challenges for China's war on water pollution.

    Science.gov (United States)

    Han, Dongmei; Currell, Matthew J; Cao, Guoliang

    2016-11-01

    China's Central government has released an ambitious plan to tackle the nation's water pollution crisis. However, this is inhibited by a lack of data, particularly for groundwater. We compiled and analyzed water quality classification data from publicly available government sources, further revealing the scale and extent of the crisis. We also compiled nitrate data in shallow and deep groundwater from a range of literature sources, covering 52 of China's groundwater systems; the most comprehensive national-scale assessment yet. Nitrate pollution at levels exceeding the US EPA's maximum contaminant level (10 mg/L NO 3 N) occurs at the 90th percentile in 25 of 36 shallow aquifers and 10 out of 37 deep or karst aquifers. Isotopic compositions of groundwater nitrate (δ 15 N and δ 18 O NO3 values ranging from -14.9‰ to 35.5‰ and -8.1‰ to 51.0‰, respectively) indicate many nitrate sources including soil nitrogen, agricultural fertilizers, untreated wastewater and/or manure, and locally show evidence of de-nitrification. From these data, it is clear that contaminated groundwater is ubiquitous in deep aquifers as well as shallow groundwater (and surface water). Deep aquifers contain water recharged tens of thousands of years before present, long before widespread anthropogenic nitrate contamination. This groundwater has therefore likely been contaminated due to rapid bypass flow along wells or other conduits. Addressing the issue of well condition is urgently needed to stop further pollution of China's deep aquifers, which are some of China's most important drinking water sources. China's new 10-point Water Pollution Plan addresses previous shortcomings, however, control and remediation of deep groundwater pollution will take decades of sustained effort. Copyright © 2016. Published by Elsevier Ltd.

  7. Boosting compound-protein interaction prediction by deep learning.

    Science.gov (United States)

    Tian, Kai; Shao, Mingyu; Wang, Yang; Guan, Jihong; Zhou, Shuigeng

    2016-11-01

    The identification of interactions between compounds and proteins plays an important role in network pharmacology and drug discovery. However, experimentally identifying compound-protein interactions (CPIs) is generally expensive and time-consuming, computational approaches are thus introduced. Among these, machine-learning based methods have achieved a considerable success. However, due to the nonlinear and imbalanced nature of biological data, many machine learning approaches have their own limitations. Recently, deep learning techniques show advantages over many state-of-the-art machine learning methods in some applications. In this study, we aim at improving the performance of CPI prediction based on deep learning, and propose a method called DL-CPI (the abbreviation of Deep Learning for Compound-Protein Interactions prediction), which employs deep neural network (DNN) to effectively learn the representations of compound-protein pairs. Extensive experiments show that DL-CPI can learn useful features of compound-protein pairs by a layerwise abstraction, and thus achieves better prediction performance than existing methods on both balanced and imbalanced datasets. Copyright © 2016 Elsevier Inc. All rights reserved.

  8. Image Captioning with Deep Bidirectional LSTMs

    OpenAIRE

    Wang, Cheng; Yang, Haojin; Bartz, Christian; Meinel, Christoph

    2016-01-01

    This work presents an end-to-end trainable deep bidirectional LSTM (Long-Short Term Memory) model for image captioning. Our model builds on a deep convolutional neural network (CNN) and two separate LSTM networks. It is capable of learning long term visual-language interactions by making use of history and future context information at high level semantic space. Two novel deep bidirectional variant models, in which we increase the depth of nonlinearity transition in different way, are propose...

  9. Deep inelastic processes and the parton model

    International Nuclear Information System (INIS)

    Altarelli, G.

    The lecture was intended as an elementary introduction to the physics of deep inelastic phenomena from the point of view of theory. General formulae and facts concerning inclusive deep inelastic processes in the form: l+N→l'+hadrons (electroproduction, neutrino scattering) are first recalled. The deep inelastic annihilation e + e - →hadrons is then envisaged. The light cone approach, the parton model and their relation are mainly emphasized

  10. Deep inelastic electron and muon scattering

    International Nuclear Information System (INIS)

    Taylor, R.E.

    1975-07-01

    From the review of deep inelastic electron and muon scattering it is concluded that the puzzle of deep inelastic scattering versus annihilation was replaced with the challenge of the new particles, that the evidence for the simplest quark-algebra models of deep inelastic processes is weaker than a year ago. Definite evidence of scale breaking was found but the specific form of that scale breaking is difficult to extract from the data. 59 references

  11. Fast, Distributed Algorithms in Deep Networks

    Science.gov (United States)

    2016-05-11

    shallow networks, additional work will need to be done in order to allow for the application of ADMM to deep nets. The ADMM method allows for quick...Quock V Le, et al. Large scale distributed deep networks. In Advances in Neural Information Processing Systems, pages 1223–1231, 2012. [11] Ken-Ichi...A TRIDENT SCHOLAR PROJECT REPORT NO. 446 Fast, Distributed Algorithms in Deep Networks by Midshipman 1/C Ryan J. Burmeister, USN

  12. Learning Transferable Features with Deep Adaptation Networks

    OpenAIRE

    Long, Mingsheng; Cao, Yue; Wang, Jianmin; Jordan, Michael I.

    2015-01-01

    Recent studies reveal that a deep neural network can learn transferable features which generalize well to novel tasks for domain adaptation. However, as deep features eventually transition from general to specific along the network, the feature transferability drops significantly in higher layers with increasing domain discrepancy. Hence, it is important to formally reduce the dataset bias and enhance the transferability in task-specific layers. In this paper, we propose a new Deep Adaptation...

  13. Velocity and stress distributions of deep seismic zone under Izu-Bonin, Japan

    Science.gov (United States)

    Jiang, Guoming; Zhang, Guibin; Jia, Zhengyuan

    2017-04-01

    Deep earthquakes can provide the deep information of the Earth directly. We have collected the waveform data from 77 deep earthquakes with depth greater than 300 km under Izu-Bonin in Japan. To obtain the velocity structures of P- and S-wave, we have inversed the double-differences of travel times from deep event-pairs. These velocity anomalies can further yield the Poisson's ratio and the porosity. Our results show that the average P-wave velocity anomaly is lower 6%, however the S-wave anomaly is higher 2% than the iasp91 model. The corresponding Poisson's ratio and porosity anomaly are -24% and -4%, respectively, which suggest that the possibility of water in the deep seismic zone is very few and the porosity might be richer. To obtain the stress distribution, we have used the ISOLA method to analyse the non-double-couple components of moment tensors of 77 deep earthquakes. The focal mechanism results show that almost half of all earthquakes have larger double-couple (DC) components, but others have clear isotropic (ISO) or compensated linear vector dipole (CLVD) components. The non-double-couple components (ISO and CLVD) seem to represent the volume around a deep earthquake changes as it occurs, which could be explained the metastable olivine phase transition. All results indicate that the metastable olivine wedge (MOW) might exist in the Pacific slab under the Izu-Bonin region and the deep earthquakes might be induced by the phase change of metastable olivine.

  14. An overview of latest deep water technologies

    International Nuclear Information System (INIS)

    Anon.

    1995-01-01

    The 8th Deep Offshore Technology Conference (DOT VIII, Rio de Janeiro, October 30 - November 3, 1995) has brought together renowned specialists in deep water development projects, as well as managers from oil companies and engineering/service companies to discuss state-of-the-art technologies and ongoing projects in the deep offshore. This paper is a compilation of the session summaries about sub sea technologies, mooring and dynamic positioning, floaters (Tension Leg Platforms (TLP) and Floating Production Storage and Off loading (FPSO)), pipelines and risers, exploration and drilling, and other deep water techniques. (J.S.)

  15. Deep learning in neural networks: an overview.

    Science.gov (United States)

    Schmidhuber, Jürgen

    2015-01-01

    In recent years, deep artificial neural networks (including recurrent ones) have won numerous contests in pattern recognition and machine learning. This historical survey compactly summarizes relevant work, much of it from the previous millennium. Shallow and Deep Learners are distinguished by the depth of their credit assignment paths, which are chains of possibly learnable, causal links between actions and effects. I review deep supervised learning (also recapitulating the history of backpropagation), unsupervised learning, reinforcement learning & evolutionary computation, and indirect search for short programs encoding deep and large networks.

  16. The deep ocean under climate change

    Science.gov (United States)

    Levin, Lisa A.; Le Bris, Nadine

    2015-11-01

    The deep ocean absorbs vast amounts of heat and carbon dioxide, providing a critical buffer to climate change but exposing vulnerable ecosystems to combined stresses of warming, ocean acidification, deoxygenation, and altered food inputs. Resulting changes may threaten biodiversity and compromise key ocean services that maintain a healthy planet and human livelihoods. There exist large gaps in understanding of the physical and ecological feedbacks that will occur. Explicit recognition of deep-ocean climate mitigation and inclusion in adaptation planning by the United Nations Framework Convention on Climate Change (UNFCCC) could help to expand deep-ocean research and observation and to protect the integrity and functions of deep-ocean ecosystems.

  17. Docker Containers for Deep Learning Experiments

    OpenAIRE

    Gerke, Paul K.

    2017-01-01

    Deep learning is a powerful tool to solve problems in the area of image analysis. The dominant compute platform for deep learning is Nvidia’s proprietary CUDA, which can only be used together with Nvidia graphics cards. The nivida-docker project allows exposing Nvidia graphics cards to docker containers and thus makes it possible to run deep learning experiments in docker containers.In our department, we use deep learning to solve problems in the area of medical image analysis and use docker ...

  18. Deep Structures of The Angola Margin

    Science.gov (United States)

    Moulin, M.; Contrucci, I.; Olivet, J.-L.; Aslanian, D.; Géli, L.; Sibuet, J.-C.

    ,4 km/s. The maximum thickness of this layer is located where the crust shows the strongest thinning at the foot of the continental slope ; and iii) a transitional domain, 160 km wide, with an average crustal thickness of 6 km. Moreover, no tilted blocks nor detachment faults are observed on the reflection seismic sections. The consequences of these observations on the models of crustal thinning classically used in the litterature are examined. Avedik, F., V. Renard, J-P. Allenou, B. Morvan, "Single bubble" air gun for deep exploration, Geophysics, 58, 366-382, 1993.

  19. Cultivating the Deep Subsurface Microbiome

    Science.gov (United States)

    Casar, C. P.; Osburn, M. R.; Flynn, T. M.; Masterson, A.; Kruger, B.

    2017-12-01

    Subterranean ecosystems are poorly understood because many microbes detected in metagenomic surveys are only distantly related to characterized isolates. Cultivating microorganisms from the deep subsurface is challenging due to its inaccessibility and potential for contamination. The Deep Mine Microbial Observatory (DeMMO) in Lead, SD however, offers access to deep microbial life via pristine fracture fluids in bedrock to a depth of 1478 m. The metabolic landscape of DeMMO was previously characterized via thermodynamic modeling coupled with genomic data, illustrating the potential for microbial inhabitants of DeMMO to utilize mineral substrates as energy sources. Here, we employ field and lab based cultivation approaches with pure minerals to link phylogeny to metabolism at DeMMO. Fracture fluids were directed through reactors filled with Fe3O4, Fe2O3, FeS2, MnO2, and FeCO3 at two sites (610 m and 1478 m) for 2 months prior to harvesting for subsequent analyses. We examined mineralogical, geochemical, and microbiological composition of the reactors via DNA sequencing, microscopy, lipid biomarker characterization, and bulk C and N isotope ratios to determine the influence of mineralogy on biofilm community development. Pre-characterized mineral chips were imaged via SEM to assay microbial growth; preliminary results suggest MnO2, Fe3O4, and Fe2O3 were most conducive to colonization. Solid materials from reactors were used as inoculum for batch cultivation experiments. Media designed to mimic fracture fluid chemistry was supplemented with mineral substrates targeting metal reducers. DNA sequences and microscopy of iron oxide-rich biofilms and fracture fluids suggest iron oxidation is a major energy source at redox transition zones where anaerobic fluids meet more oxidizing conditions. We utilized these biofilms and fluids as inoculum in gradient cultivation experiments targeting microaerophilic iron oxidizers. Cultivation of microbes endemic to DeMMO, a system

  20. Multiagent cooperation and competition with deep reinforcement learning.

    Directory of Open Access Journals (Sweden)

    Ardi Tampuu

    Full Text Available Evolution of cooperation and competition can appear when multiple adaptive agents share a biological, social, or technological niche. In the present work we study how cooperation and competition emerge between autonomous agents that learn by reinforcement while using only their raw visual input as the state representation. In particular, we extend the Deep Q-Learning framework to multiagent environments to investigate the interaction between two learning agents in the well-known video game Pong. By manipulating the classical rewarding scheme of Pong we show how competitive and collaborative behaviors emerge. We also describe the progression from competitive to collaborative behavior when the incentive to cooperate is increased. Finally we show how learning by playing against another adaptive agent, instead of against a hard-wired algorithm, results in more robust strategies. The present work shows that Deep Q-Networks can become a useful tool for studying decentralized learning of multiagent systems coping with high-dimensional environments.

  1. Multiagent cooperation and competition with deep reinforcement learning.

    Science.gov (United States)

    Tampuu, Ardi; Matiisen, Tambet; Kodelja, Dorian; Kuzovkin, Ilya; Korjus, Kristjan; Aru, Juhan; Aru, Jaan; Vicente, Raul

    2017-01-01

    Evolution of cooperation and competition can appear when multiple adaptive agents share a biological, social, or technological niche. In the present work we study how cooperation and competition emerge between autonomous agents that learn by reinforcement while using only their raw visual input as the state representation. In particular, we extend the Deep Q-Learning framework to multiagent environments to investigate the interaction between two learning agents in the well-known video game Pong. By manipulating the classical rewarding scheme of Pong we show how competitive and collaborative behaviors emerge. We also describe the progression from competitive to collaborative behavior when the incentive to cooperate is increased. Finally we show how learning by playing against another adaptive agent, instead of against a hard-wired algorithm, results in more robust strategies. The present work shows that Deep Q-Networks can become a useful tool for studying decentralized learning of multiagent systems coping with high-dimensional environments.

  2. Multiagent cooperation and competition with deep reinforcement learning

    Science.gov (United States)

    Kodelja, Dorian; Kuzovkin, Ilya; Korjus, Kristjan; Aru, Juhan; Aru, Jaan; Vicente, Raul

    2017-01-01

    Evolution of cooperation and competition can appear when multiple adaptive agents share a biological, social, or technological niche. In the present work we study how cooperation and competition emerge between autonomous agents that learn by reinforcement while using only their raw visual input as the state representation. In particular, we extend the Deep Q-Learning framework to multiagent environments to investigate the interaction between two learning agents in the well-known video game Pong. By manipulating the classical rewarding scheme of Pong we show how competitive and collaborative behaviors emerge. We also describe the progression from competitive to collaborative behavior when the incentive to cooperate is increased. Finally we show how learning by playing against another adaptive agent, instead of against a hard-wired algorithm, results in more robust strategies. The present work shows that Deep Q-Networks can become a useful tool for studying decentralized learning of multiagent systems coping with high-dimensional environments. PMID:28380078

  3. Modeling of Antenna for Deep Target Hydrocarbon Exploration

    Directory of Open Access Journals (Sweden)

    Nadeem Nasir

    2017-11-01

    Full Text Available Nowadays control source electromagnetic method is used for offshore hydrocarbon exploration. Hydrocarbon detection in sea bed logging (SBL is a very challenging task for deep target hydrocarbon reservoir. Response of electromagnetic (EM field from marine environment is very low and it is very difficult to predict deep target reservoir below 2km from the sea floor. This work premise deals with modeling of new antenna for deep water deep target hydrocarbon exploration. Conventional and new EM antennas at 0.125Hz frequency are used in modeling for the detection of deep target hydrocarbon  reservoir.  The  proposed  area  of  the  seabed model   (40km ´ 40km   was   simulated   by using CST (computer simulation technology EM studio based on Finite Integration Method (FIM. Electromagnetic field components were compared at 500m target depth and it was concluded that Ex and Hz components shows better resistivity contrast. Comparison of conventional and new antenna for different target  depths  was  done in  our  proposed  model.  From  the results, it was observed that conventional antenna at 0.125Hz shows 70% ,86% resistivity contrast at target depth of 1000m where   as   new   antenna   showed   329%, 355%   resistivity contrast at the same target depth for Ex and Hz field respectively.  It  was  also  investigated  that  at  frequency of0.125Hz, new antenna gave 46% better delineation of hydrocarbon at 4000m target depth. This is due to focusing of electromagnetic waves by using new antenna. New antenna design gave 125% more extra depth than straight antenna for deep target hydrocarbon detection. Numerical modeling for straight  and  new antenna  was also done to know general equation for electromagnetic field behavior with target depth. From this numerical model it was speculated that this new antenna can detect up to 4.5 km target depth. This new EM antenna may open new frontiers for oil and gas

  4. Unsupervised deep learning reveals prognostically relevant subtypes of glioblastoma.

    Science.gov (United States)

    Young, Jonathan D; Cai, Chunhui; Lu, Xinghua

    2017-10-03

    further investigate the disease mechanisms underlying each of these clusters. In summary, we show that a deep learning model can be trained to represent biologically and clinically meaningful abstractions of cancer gene expression data. Understanding what additional relationships these hidden layer abstractions have with the cancer cellular signaling system could have a significant impact on the understanding and treatment of cancer.

  5. Deep inelastic scattering and disquarks

    International Nuclear Information System (INIS)

    Anselmino, M.

    1993-01-01

    The most comprehensive and detailed analyses of the existing data on the structure function F 2 (x, Q 2 ) of free nucleons, from the deep inelastic scattering (DIS) of charged leptons on hydrogen and deuterium targets, have proved beyond any doubt that higher twist, 1/Q 2 corrections are needed in order to obtain a perfect agreement between perturbative QCD predictions and the data. These higher twist corrections take into account two quark correlations inside the nucleon; it is then natural to try to model them in the quark-diquark model of the proton. In so doing all interactions between the two quarks inside the diquark, both perturbative and non perturbative, are supposed to be taken into account. (orig./HSI)

  6. Detector for deep well logging

    International Nuclear Information System (INIS)

    1976-01-01

    A substantial improvement in the useful life and efficiency of a deep-well scintillation detector is achieved by a unique construction wherein the steel cylinder enclosing the sodium iodide scintillation crystal is provided with a tapered recess to receive a glass window which has a high transmittance at the critical wavelength and, for glass, a high coefficient of thermal expansion. A special high-temperature epoxy adhesive composition is employed to form a relatively thick sealing annulus which keeps the glass window in the tapered recess and compensates for the differences in coefficients of expansion between the container and glass so as to maintain a hermetic seal as the unit is subjected to a wide range of temperature

  7. Deep borehole disposal of plutonium

    International Nuclear Information System (INIS)

    Gibb, F. G. F.; Taylor, K. J.; Burakov, B. E.

    2008-01-01

    Excess plutonium not destined for burning as MOX or in Generation IV reactors is both a long-term waste management problem and a security threat. Immobilisation in mineral and ceramic-based waste forms for interim safe storage and eventual disposal is a widely proposed first step. The safest and most secure form of geological disposal for Pu yet suggested is in very deep boreholes and we propose here that the key to successful combination of these immobilisation and disposal concepts is the encapsulation of the waste form in small cylinders of recrystallized granite. The underlying science is discussed and the results of high pressure and temperature experiments on zircon, depleted UO 2 and Ce-doped cubic zirconia enclosed in granitic melts are presented. The outcomes of these experiments demonstrate the viability of the proposed solution and that Pu could be successfully isolated from its environment for many millions of years. (authors)

  8. Automatic Differentiation and Deep Learning

    CERN Multimedia

    CERN. Geneva

    2018-01-01

    Statistical learning has been getting more and more interest from the particle-physics community in recent times, with neural networks and gradient-based optimization being a focus. In this talk we shall discuss three things: automatic differention tools: tools to quickly build DAGs of computation that are fully differentiable. We shall focus on one such tool "PyTorch".  Easy deployment of trained neural networks into large systems with many constraints: for example, deploying a model at the reconstruction phase where the neural network has to be integrated into CERN's bulk data-processing C++-only environment Some recent models in deep learning for segmentation and generation that might be useful for particle physics problems.

  9. Jets in deep inelastic scattering

    International Nuclear Information System (INIS)

    Joensson, L.

    1995-01-01

    Jet production in deep inelastic scattering provides a basis for the investigation of various phenomena related to QCD. Two-jet production at large Q 2 has been studied and the distributions with respect to the partonic scaling variables have been compared to models and to next to leading order calculations. The first observations of azimuthal asymmetries of jets produced in first order α s processes have been obtained. The gluon initiated boson-gluon fusion process permits a direct determination of the gluon density of the proton from an analysis of the jets produced in the hard scattering process. A comparison of these results with those from indirect extractions of the gluon density provides an important test of QCD. (author)

  10. NESTOR Deep Sea Neutrino Telescope

    International Nuclear Information System (INIS)

    Aggouras, G.; Anassontzis, E.G.; Ball, A.E.; Bourlis, G.; Chinowsky, W.; Fahrun, E.; Grammatikakis, G.; Green, C.; Grieder, P.; Katrivanos, P.; Koske, P.; Leisos, A.; Markopoulos, E.; Minkowsky, P.; Nygren, D.; Papageorgiou, K.; Przybylski, G.; Resvanis, L.K.; Siotis, I.; Sopher, J.; Staveris-Polikalas, A.; Tsagli, V.; Tsirigotis, A.; Tzamarias, S.; Zhukov, V.A.

    2006-01-01

    One module of NESTOR, the Mediterranean deep-sea neutrino telescope, was deployed at a depth of 4000m, 14km off the Sapienza Island, off the South West coast of Greece. The deployment site provides excellent environmental characteristics. The deployed NESTOR module is constructed as a hexagonal star like latticed titanium star with 12 Optical Modules and an one-meter diameter titanium sphere which houses the electronics. Power and data were transferred through a 30km electro-optical cable to the shore laboratory. In this report we describe briefly the detector and the detector electronics and discuss the first physics data acquired and give the zenith angular distribution of the reconstructed muons

  11. Deep Borehole Disposal Safety Analysis.

    Energy Technology Data Exchange (ETDEWEB)

    Freeze, Geoffrey A. [Sandia National Laboratories (SNL-NM), Albuquerque, NM (United States); Stein, Emily [Sandia National Laboratories (SNL-NM), Albuquerque, NM (United States); Price, Laura L. [Sandia National Laboratories (SNL-NM), Albuquerque, NM (United States); MacKinnon, Robert J. [Sandia National Laboratories (SNL-NM), Albuquerque, NM (United States); Tillman, Jack Bruce [Sandia National Laboratories (SNL-NM), Albuquerque, NM (United States)

    2016-10-01

    This report presents a preliminary safety analysis for the deep borehole disposal (DBD) concept, using a safety case framework. A safety case is an integrated collection of qualitative and quantitative arguments, evidence, and analyses that substantiate the safety, and the level of confidence in the safety, of a geologic repository. This safety case framework for DBD follows the outline of the elements of a safety case, and identifies the types of information that will be required to satisfy these elements. At this very preliminary phase of development, the DBD safety case focuses on the generic feasibility of the DBD concept. It is based on potential system designs, waste forms, engineering, and geologic conditions; however, no specific site or regulatory framework exists. It will progress to a site-specific safety case as the DBD concept advances into a site-specific phase, progressing through consent-based site selection and site investigation and characterization.

  12. Corium quench in deep pool mixing experiments

    International Nuclear Information System (INIS)

    Spencer, B.W.; McUmber, L.; Gregorash, D.; Aeschlimann, R.; Sienicki, J.J.

    1985-01-01

    The results of two recent corium-water thermal interaction (CWTI) tests are described in which a stream of molten corium was poured into a deep pool of water in order to determine the mixing behavior, the corium-to-water heat transfer rates, and the characteristic sizes of the quenched debris. The corium composition was 60% UO 2 , 16% ZrO 2 , and 24% stainless steel by weight; its initial temperature was 3080 K, approx.160 K above the oxide phase liquidus temperature. The corium pour stream was a single-phase 2.2 cm dia liquid column which entered the water pool in film boiling at approx.4 m/s. The water subcooling was 6 and 75C in the two tests. Test results showed that with low subcooling, rapid steam generation caused the pool to boil up into a high void fraction regime. In contrast, with large subcooling no net steam generation occurred, and the pool remained relatively quiescent. Breakup of the jet appeared to occur by surface stripping. In neither test was the breakup complete during transit through the 32 cm deep water pool, and molten corium channeled to the base where it formed a melt layer. The characteristic heat transfer rates measured 3.5 MJ/s and 2.7 MJ/s during the fall stage for small and large subcooling, respectively; during the initial stage of bed quench, the surface heat fluxes measured 2.4 MW/m 2 and 3.7 MW/m 2 , respectively. A small mass of particles was formed in each test, measuring typically 0.1 to 1 mm and 1 to 5 mm dia for the large and small subcooling conditions, respectively. 9 refs., 13 figs., 1 tab

  13. Deep Recurrent Neural Networks for Supernovae Classification

    Science.gov (United States)

    Charnock, Tom; Moss, Adam

    2017-03-01

    We apply deep recurrent neural networks, which are capable of learning complex sequential information, to classify supernovae (code available at https://github.com/adammoss/supernovae). The observational time and filter fluxes are used as inputs to the network, but since the inputs are agnostic, additional data such as host galaxy information can also be included. Using the Supernovae Photometric Classification Challenge (SPCC) data, we find that deep networks are capable of learning about light curves, however the performance of the network is highly sensitive to the amount of training data. For a training size of 50% of the representational SPCC data set (around 104 supernovae) we obtain a type-Ia versus non-type-Ia classification accuracy of 94.7%, an area under the Receiver Operating Characteristic curve AUC of 0.986 and an SPCC figure-of-merit F 1 = 0.64. When using only the data for the early-epoch challenge defined by the SPCC, we achieve a classification accuracy of 93.1%, AUC of 0.977, and F 1 = 0.58, results almost as good as with the whole light curve. By employing bidirectional neural networks, we can acquire impressive classification results between supernovae types I, II and III at an accuracy of 90.4% and AUC of 0.974. We also apply a pre-trained model to obtain classification probabilities as a function of time and show that it can give early indications of supernovae type. Our method is competitive with existing algorithms and has applications for future large-scale photometric surveys.

  14. Can bats sense smoke during deep torpor?

    Science.gov (United States)

    Doty, Anna C; Currie, Shannon E; Stawski, Clare; Geiser, Fritz

    2018-03-01

    While torpor is a beneficial energy-saving strategy, it may incur costs if an animal is unable to respond appropriately to external stimuli, which is particularly true when it is necessary to escape from threats such as fire. We aimed to determine whether torpid bats, which are potentially threatened because they must fly to escape, can sense smoke and whether respiration rate (RR), heart rate (HR) and reaction time of torpid bats prior to and following smoke introduction is temperature-dependent. To test this we quantified RR and HR of captive Australian tree-roosting bats, Nyctophilus gouldi (n=5, ~10g), in steady-state torpor in response to short-term exposure to smoke from Eucalyptus spp. leaves between ambient temperatures (T a ) of 11 and 23°C. Bats at lower T a took significantly longer (28-fold) to respond to smoke, indicated by a cessation of episodic breathing and a rapid increase in RR. Bats at lower T a returned to torpor more swiftly following smoke exposure than bats at higher T a . Interestingly, bats at T a <15°C never returned to thermoconforming steady-state torpor prior to the end of the experimental day, whereas all bats at T a ≥15°C did, as indicated by apnoeic HR. This shows that although bats at lower T a took longer to respond, they appear to maintain vigilance and prevent deep torpor after the first smoke exposure, likely to enable fast escape. Our study reveals that bats can respond to smoke stimuli while in deep torpor. These results are particularly vital within the framework of fire management conducted at T a <15°C, as most management burns are undertaken during winter when bats will likely respond more slowly to fire cues such as smoke, delaying the time to escape from the fire. Copyright © 2017 Elsevier Inc. All rights reserved.

  15. Deep Learning and Applications in Computational Biology

    KAUST Repository

    Zeng, Jianyang

    2016-01-26

    RNA-binding proteins (RBPs) play important roles in the post-transcriptional control of RNAs. Identifying RBP binding sites and characterizing RBP binding preferences are key steps toward understanding the basic mechanisms of the post-transcriptional gene regulation. Though numerous computational methods have been developed for modeling RBP binding preferences, discovering a complete structural representation of the RBP targets by integrating their available structural features in all three dimensions is still a challenging task. In this work, we develop a general and flexible deep learning framework for modeling structural binding preferences and predicting binding sites of RBPs, which takes (predicted) RNA tertiary structural information into account for the first time. Our framework constructs a unified representation that characterizes the structural specificities of RBP targets in all three dimensions, which can be further used to predict novel candidate binding sites and discover potential binding motifs. Through testing on the real CLIP-seq datasets, we have demonstrated that our deep learning framework can automatically extract effective hidden structural features from the encoded raw sequence and structural profiles, and predict accurate RBP binding sites. In addition, we have conducted the first study to show that integrating the additional RNA tertiary structural features can improve the model performance in predicting RBP binding sites, especially for the polypyrimidine tract-binding protein (PTB), which also provides a new evidence to support the view that RBPs may own specific tertiary structural binding preferences. In particular, the tests on the internal ribosome entry site (IRES) segments yield satisfiable results with experimental support from the literature and further demonstrate the necessity of incorporating RNA tertiary structural information into the prediction model. The source code of our approach can be found in https://github.com/thucombio/deepnet-rbp.

  16. Dimensioning of lining galleries in deep clays

    International Nuclear Information System (INIS)

    Bernaud, D.; Rousset, G.

    1991-01-01

    The aim of the work presented in this report is to study the mechanical behaviour of lining galleries in deep clays. This text constitutes a part of the researches on the feasibility of a geological disposal of radioactive waste, which the scope is to assure the gallery long term stabilization and also to optimize its dimensioning. In particular, we are interested here in the study of a closure controlled lining, that constitutes a direct application of the convergence-confinement method, especially well fitted to deep clays. The presentation and interpretation of the convergence controlled lining test, which was performed in the experimental gallery of Mol in Belgium, is given in this report. The instrumentation was conceived in order to find out the stress field exerced by the rockmass on the lining, the internal stress field inside the lining and the gallery closure. The analysis of all measurements results, obtained between november 1987 and December 1989, shows that they are all in good agreement and that the lining design was well chosen. Two years after the gallery construction, the average closure is of the order of 2% and the average confinement pressure is about 1.6 MPa (the third of the lithostatic pressure). The time dependent effects of the rockmass are very well modelled by the non linear elasto-viscoplastic law developed at L.M.S. with the laboratory tests. The elastic-plastic model of the lining are shown to be well fitted to simulate the sliding of the ribs. Finally, the numerical results have shown a very good agreement with the measurements results

  17. Deep Question Answering for protein annotation.

    Science.gov (United States)

    Gobeill, Julien; Gaudinat, Arnaud; Pasche, Emilie; Vishnyakova, Dina; Gaudet, Pascale; Bairoch, Amos; Ruch, Patrick

    2015-01-01

    Biomedical professionals have access to a huge amount of literature, but when they use a search engine, they often have to deal with too many documents to efficiently find the appropriate information in a reasonable time. In this perspective, question-answering (QA) engines are designed to display answers, which were automatically extracted from the retrieved documents. Standard QA engines in literature process a user question, then retrieve relevant documents and finally extract some possible answers out of these documents using various named-entity recognition processes. In our study, we try to answer complex genomics questions, which can be adequately answered only using Gene Ontology (GO) concepts. Such complex answers cannot be found using state-of-the-art dictionary- and redundancy-based QA engines. We compare the effectiveness of two dictionary-based classifiers for extracting correct GO answers from a large set of 100 retrieved abstracts per question. In the same way, we also investigate the power of GOCat, a GO supervised classifier. GOCat exploits the GOA database to propose GO concepts that were annotated by curators for similar abstracts. This approach is called deep QA, as it adds an original classification step, and exploits curated biological data to infer answers, which are not explicitly mentioned in the retrieved documents. We show that for complex answers such as protein functional descriptions, the redundancy phenomenon has a limited effect. Similarly usual dictionary-based approaches are relatively ineffective. In contrast, we demonstrate how existing curated data, beyond information extraction, can be exploited by a supervised classifier, such as GOCat, to massively improve both the quantity and the quality of the answers with a +100% improvement for both recall and precision. Database URL: http://eagl.unige.ch/DeepQA4PA/. © The Author(s) 2015. Published by Oxford University Press.

  18. Lung Nodule Detection via Deep Reinforcement Learning

    Directory of Open Access Journals (Sweden)

    Issa Ali

    2018-04-01

    Full Text Available Lung cancer is the most common cause of cancer-related death globally. As a preventive measure, the United States Preventive Services Task Force (USPSTF recommends annual screening of high risk individuals with low-dose computed tomography (CT. The resulting volume of CT scans from millions of people will pose a significant challenge for radiologists to interpret. To fill this gap, computer-aided detection (CAD algorithms may prove to be the most promising solution. A crucial first step in the analysis of lung cancer screening results using CAD is the detection of pulmonary nodules, which may represent early-stage lung cancer. The objective of this work is to develop and validate a reinforcement learning model based on deep artificial neural networks for early detection of lung nodules in thoracic CT images. Inspired by the AlphaGo system, our deep learning algorithm takes a raw CT image as input and views it as a collection of states, and output a classification of whether a nodule is present or not. The dataset used to train our model is the LIDC/IDRI database hosted by the lung nodule analysis (LUNA challenge. In total, there are 888 CT scans with annotations based on agreement from at least three out of four radiologists. As a result, there are 590 individuals having one or more nodules, and 298 having none. Our training results yielded an overall accuracy of 99.1% [sensitivity 99.2%, specificity 99.1%, positive predictive value (PPV 99.1%, negative predictive value (NPV 99.2%]. In our test, the results yielded an overall accuracy of 64.4% (sensitivity 58.9%, specificity 55.3%, PPV 54.2%, and NPV 60.0%. These early results show promise in solving the major issue of false positives in CT screening of lung nodules, and may help to save unnecessary follow-up tests and expenditures.

  19. Deep neural mapping support vector machines.

    Science.gov (United States)

    Li, Yujian; Zhang, Ting

    2017-09-01

    The choice of kernel has an important effect on the performance of a support vector machine (SVM). The effect could be reduced by NEUROSVM, an architecture using multilayer perceptron for feature extraction and SVM for classification. In binary classification, a general linear kernel NEUROSVM can be theoretically simplified as an input layer, many hidden layers, and an SVM output layer. As a feature extractor, the sub-network composed of the input and hidden layers is first trained together with a virtual ordinary output layer by backpropagation, then with the output of its last hidden layer taken as input of the SVM classifier for further training separately. By taking the sub-network as a kernel mapping from the original input space into a feature space, we present a novel model, called deep neural mapping support vector machine (DNMSVM), from the viewpoint of deep learning. This model is also a new and general kernel learning method, where the kernel mapping is indeed an explicit function expressed as a sub-network, different from an implicit function induced by a kernel function traditionally. Moreover, we exploit a two-stage procedure of contrastive divergence learning and gradient descent for DNMSVM to jointly training an adaptive kernel mapping instead of a kernel function, without requirement of kernel tricks. As a whole of the sub-network and the SVM classifier, the joint training of DNMSVM is done by using gradient descent to optimize the objective function with the sub-network layer-wise pre-trained via contrastive divergence learning of restricted Boltzmann machines. Compared to the separate training of NEUROSVM, the joint training is a new algorithm for DNMSVM to have advantages over NEUROSVM. Experimental results show that DNMSVM can outperform NEUROSVM and RBFSVM (i.e., SVM with the kernel of radial basis function), demonstrating its effectiveness. Copyright © 2017 Elsevier Ltd. All rights reserved.

  20. DeepRT: deep learning for peptide retention time prediction in proteomics

    OpenAIRE

    Ma, Chunwei; Zhu, Zhiyong; Ye, Jun; Yang, Jiarui; Pei, Jianguo; Xu, Shaohang; Zhou, Ruo; Yu, Chang; Mo, Fan; Wen, Bo; Liu, Siqi

    2017-01-01

    Accurate predictions of peptide retention times (RT) in liquid chromatography have many applications in mass spectrometry-based proteomics. Herein, we present DeepRT, a deep learning based software for peptide retention time prediction. DeepRT automatically learns features directly from the peptide sequences using the deep convolutional Neural Network (CNN) and Recurrent Neural Network (RNN) model, which eliminates the need to use hand-crafted features or rules. After the feature learning, pr...

  1. First LOCSMITH locations of deep moonquakes

    Science.gov (United States)

    Hempel, S.; Knapmeyer, M.; Sens-Schönfelder, C.; Oberst, J.

    2008-09-01

    Introduction Several thousand seismic events were recorded by the Apollo seismic network from 19691977. Different types of events can be distinguished: meteoroid impacts, thermal quakes and internally caused moonquakes. The latter subdivide into shallow (100 to 300km) and deep moonquakes (700 to 1100km), which are by far the most common events. The deep quakes would be no immediate danger to inhabitated stations on the Earth's Moon because of their relatively low magnitude and great depth. However, they bear important information on lunar structure and evolution, and their distribution probably reflects their source mechanism. In this study, we reinvestigate location patterns of deep lunar quakes. LOCSMITH The core of this study is a new location method (LOCSMITH, [1]). This algorithm uses time intervals rather than time instants as input, which contain the dedicated arrival with probability 1. LOCSMITH models and compares theoretical and actual travel times on a global scale and uses an adaptive grid to search source locations compatible with all observations. The output is a set of all possible hypocenters for the considered region of repeating, tidally triggered moonquake activity, called clusters. The shape and size of these sets gives a better estimate of the location uncertainty than the formal standard deviations returned by classical methods. This is used for grading of deep moonquake clusters according to the currently available data quality. Classification of deep moonquakes As first step, we establish a reciprocal dependence of size and shape of LOCSMITH location clouds on number of arrivals. Four different shapes are recognized, listed here in an order corresponding to decreasing spatial resolution: 1. "Balls", which are well defined and relatively small types of sets resembling the commonly assumed error ellipsoid. These are found in the best cases with many observations. Locations in this shape are obtained for clusters 1, 18 or 33, these were already

  2. The influence of deep cryogenic treatment on the properties of high-vanadium alloy steel

    Energy Technology Data Exchange (ETDEWEB)

    Li, Haizhi [Key Laboratory of Electromagnetic Processing of Materials (Ministry of Education), Northeastern University, Shenyang 110819 (China); Tong, Weiping, E-mail: wptong@mail.neu.edu.cn [Key Laboratory of Electromagnetic Processing of Materials (Ministry of Education), Northeastern University, Shenyang 110819 (China); Cui, Junjun [State Key Laboratory of Rolling and Automation, Northeastern University, Shenyang 110819 (China); Zhang, Hui [Key Laboratory of Electromagnetic Processing of Materials (Ministry of Education), Northeastern University, Shenyang 110819 (China); Chen, Liqing [State Key Laboratory of Rolling and Automation, Northeastern University, Shenyang 110819 (China); Zuo, Liang [Key Laboratory for Anisotropy and Texture of Materials (Ministry of Education), School of Materials and Metallurgy, Northeastern University, Shenyang 110819 (China)

    2016-04-26

    Deep cryogenic treatment can improve the mechanical properties of many metallic materials, but there are few reports of the effect of deep cryogenic treatment on high-vanadium alloy steel. The main objective of this work is to investigate the effect of deep cryogenic treatment on the microstructure, hardness, impact toughness and abrasive wear resistance of high-vanadium alloy steel. The results show that large amounts of small secondary carbide precipitation after deep cryogenic treatment and microcracks were detected and occurred preferentially at carbide/matrix interfaces; except for the hardness, the mechanical properties increased compared to those of the conventional treatment sample. By increasing the deep cryogenic processing time and cycle number, impact toughness and abrasive wear resistance can be further improved, the carbide contents continuously increase, and the hardness decreases.

  3. A deep auto-encoder model for gene expression prediction.

    Science.gov (United States)

    Xie, Rui; Wen, Jia; Quitadamo, Andrew; Cheng, Jianlin; Shi, Xinghua

    2017-11-17

    Gene expression is a key intermediate level that genotypes lead to a particular trait. Gene expression is affected by various factors including genotypes of genetic variants. With an aim of delineating the genetic impact on gene expression, we build a deep auto-encoder model to assess how good genetic variants will contribute to gene expression changes. This new deep learning model is a regression-based predictive model based on the MultiLayer Perceptron and Stacked Denoising Auto-encoder (MLP-SAE). The model is trained using a stacked denoising auto-encoder for feature selection and a multilayer perceptron framework for backpropagation. We further improve the model by introducing dropout to prevent overfitting and improve performance. To demonstrate the usage of this model, we apply MLP-SAE to a real genomic datasets with genotypes and gene expression profiles measured in yeast. Our results show that the MLP-SAE model with dropout outperforms other models including Lasso, Random Forests and the MLP-SAE model without dropout. Using the MLP-SAE model with dropout, we show that gene expression quantifications predicted by the model solely based on genotypes, align well with true gene expression patterns. We provide a deep auto-encoder model for predicting gene expression from SNP genotypes. This study demonstrates that deep learning is appropriate for tackling another genomic problem, i.e., building predictive models to understand genotypes' contribution to gene expression. With the emerging availability of richer genomic data, we anticipate that deep learning models play a bigger role in modeling and interpreting genomics.

  4. Global diversity and biogeography of deep-sea pelagic prokaryotes

    KAUST Repository

    Salazar, Guillem

    2015-08-07

    The deep-sea is the largest biome of the biosphere, and contains more than half of the whole ocean\\'s microbes. Uncovering their general patterns of diversity and community structure at a global scale remains a great challenge, as only fragmentary information of deep-sea microbial diversity exists based on regional-scale studies. Here we report the first globally comprehensive survey of the prokaryotic communities inhabiting the bathypelagic ocean using high-throughput sequencing of the 16S rRNA gene. This work identifies the dominant prokaryotes in the pelagic deep ocean and reveals that 50% of the operational taxonomic units (OTUs) belong to previously unknown prokaryotic taxa, most of which are rare and appear in just a few samples. We show that whereas the local richness of communities is comparable to that observed in previous regional studies, the global pool of prokaryotic taxa detected is modest (∼3600 OTUs), as a high proportion of OTUs are shared among samples. The water masses appear to act as clear drivers of the geographical distribution of both particle-attached and free-living prokaryotes. In addition, we show that the deep-oceanic basins in which the bathypelagic realm is divided contain different particle-attached (but not free-living) microbial communities. The combination of the aging of the water masses and a lack of complete dispersal are identified as the main drivers for this biogeographical pattern. All together, we identify the potential of the deep ocean as a reservoir of still unknown biological diversity with a higher degree of spatial complexity than hitherto considered.

  5. Global diversity and biogeography of deep-sea pelagic prokaryotes

    KAUST Repository

    Salazar, Guillem; Cornejo-Castillo, Francisco M.; Bení tez-Barrios, Veró nica; Fraile-Nuez, Eugenio; Á lvarez-Salgado, X. Antó n; Duarte, Carlos M.; Gasol, Josep M.; Acinas, Silvia G.

    2015-01-01

    The deep-sea is the largest biome of the biosphere, and contains more than half of the whole ocean's microbes. Uncovering their general patterns of diversity and community structure at a global scale remains a great challenge, as only fragmentary information of deep-sea microbial diversity exists based on regional-scale studies. Here we report the first globally comprehensive survey of the prokaryotic communities inhabiting the bathypelagic ocean using high-throughput sequencing of the 16S rRNA gene. This work identifies the dominant prokaryotes in the pelagic deep ocean and reveals that 50% of the operational taxonomic units (OTUs) belong to previously unknown prokaryotic taxa, most of which are rare and appear in just a few samples. We show that whereas the local richness of communities is comparable to that observed in previous regional studies, the global pool of prokaryotic taxa detected is modest (∼3600 OTUs), as a high proportion of OTUs are shared among samples. The water masses appear to act as clear drivers of the geographical distribution of both particle-attached and free-living prokaryotes. In addition, we show that the deep-oceanic basins in which the bathypelagic realm is divided contain different particle-attached (but not free-living) microbial communities. The combination of the aging of the water masses and a lack of complete dispersal are identified as the main drivers for this biogeographical pattern. All together, we identify the potential of the deep ocean as a reservoir of still unknown biological diversity with a higher degree of spatial complexity than hitherto considered.

  6. Show Horse Welfare: Horse Show Competitors' Understanding, Awareness, and Perceptions of Equine Welfare.

    Science.gov (United States)

    Voigt, Melissa A; Hiney, Kristina; Richardson, Jennifer C; Waite, Karen; Borron, Abigail; Brady, Colleen M

    2016-01-01

    The purpose of this study was to gain a better understanding of stock-type horse show competitors' understanding of welfare and level of concern for stock-type show horses' welfare. Data were collected through an online questionnaire that included questions relating to (a) interest and general understanding of horse welfare, (b) welfare concerns of the horse show industry and specifically the stock-type horse show industry, (c) decision-making influences, and (d) level of empathic characteristics. The majority of respondents indicated they agree or strongly agree that physical metrics should be a factor when assessing horse welfare, while fewer agreed that behavioral and mental metrics should be a factor. Respondent empathy levels were moderate to high and were positively correlated with the belief that mental and behavioral metrics should be a factor in assessing horse welfare. Respondents indicated the inhumane practices that most often occur at stock-type shows include excessive jerking on reins, excessive spurring, and induced excessive unnatural movement. Additionally, respondents indicated association rules, hired trainers, and hired riding instructors are the most influential regarding the decisions they make related to their horses' care and treatment.

  7. Controls on deep drainage beneath the root soil zone in snowmelt-dominated environments

    Science.gov (United States)

    Hammond, J. C.; Harpold, A. A.; Kampf, S. K.

    2017-12-01

    Snowmelt is the dominant source of streamflow generation and groundwater recharge in many high elevation and high latitude locations, yet we still lack a detailed understanding of how snowmelt is partitioned between the soil, deep drainage, and streamflow under a variety of soil, climate, and snow conditions. Here we use Hydrus 1-D simulations with historical inputs from five SNOTEL snow monitoring sites in each of three regions, Cascades, Sierra, and Southern Rockies, to investigate how inter-annual variability on water input rate and duration affects soil saturation and deep drainage. Each input scenario was run with three different soil profiles of varying hydraulic conductivity, soil texture, and bulk density. We also created artificial snowmelt scenarios to test how snowmelt intermittence affects deep drainage. Results indicate that precipitation is the strongest predictor (R2 = 0.83) of deep drainage below the root zone, with weaker relationships observed between deep drainage and snow persistence, peak snow water equivalent, and melt rate. The ratio of deep drainage to precipitation shows a stronger positive relationship to melt rate suggesting that a greater fraction of input becomes deep drainage at higher melt rates. For a given amount of precipitation, rapid, concentrated snowmelt may create greater deep drainage below the root zone than slower, intermittent melt. Deep drainage requires saturation below the root zone, so saturated hydraulic conductivity serves as a primary control on deep drainage magnitude. Deep drainage response to climate is mostly independent of soil texture because of its reliance on saturated conditions. Mean water year saturations of deep soil layers can predict deep drainage and may be a useful way to compare sites in soils with soil hydraulic porosities. The unit depth of surface runoff often is often greater than deep drainage at daily and annual timescales, as snowmelt exceeds infiltration capacity in near-surface soil layers

  8. Deep Recurrent Neural Networks for Human Activity Recognition

    Directory of Open Access Journals (Sweden)

    Abdulmajid Murad

    2017-11-01

    Full Text Available Adopting deep learning methods for human activity recognition has been effective in extracting discriminative features from raw input sequences acquired from body-worn sensors. Although human movements are encoded in a sequence of successive samples in time, typical machine learning methods perform recognition tasks without exploiting the temporal correlations between input data samples. Convolutional neural networks (CNNs address this issue by using convolutions across a one-dimensional temporal sequence to capture dependencies among input data. However, the size of convolutional kernels restricts the captured range of dependencies between data samples. As a result, typical models are unadaptable to a wide range of activity-recognition configurations and require fixed-length input windows. In this paper, we propose the use of deep recurrent neural networks (DRNNs for building recognition models that are capable of capturing long-range dependencies in variable-length input sequences. We present unidirectional, bidirectional, and cascaded architectures based on long short-term memory (LSTM DRNNs and evaluate their effectiveness on miscellaneous benchmark datasets. Experimental results show that our proposed models outperform methods employing conventional machine learning, such as support vector machine (SVM and k-nearest neighbors (KNN. Additionally, the proposed models yield better performance than other deep learning techniques, such as deep believe networks (DBNs and CNNs.

  9. Deep learning-based Diabetic Retinopathy assessment on embedded system.

    Science.gov (United States)

    Ardiyanto, Igi; Nugroho, Hanung Adi; Buana, Ratna Lestari Budiani

    2017-07-01

    Diabetic Retinopathy (DR) is a disease which affect the vision ability. The observation by an ophthalmologist usually conducted by analyzing the retinal images of the patient which are marked by some DR features. However some misdiagnosis are usually found due to human error. Here, a deep learning-based low-cost embedded system is established to assist the doctor for grading the severity of the DR from the retinal images. A compact deep learning algorithm named Deep-DR-Net which fits on a small embedded board is afterwards proposed for such purposes. In the heart of Deep-DR-Net, a cascaded encoder-classifier network is arranged using residual style for ensuring the small model size. The usage of different types of convolutional layers subsequently guarantees the features richness of the network for differentiating the grade of the DR. Experimental results show the capability of the proposed system for detecting the existence as well as grading the severity of the DR symptomps.

  10. Active Cooling of Oil after Deep-frying.

    Science.gov (United States)

    Totani, Nagao; Yasaki, Naoko; Doi, Rena; Hasegawa, Etsuko

    2017-10-01

    Oil used for deep-frying is often left to stand after cooking. A major concern is oxidation during standing that might be avoidable, especially in the case of oil used repeatedly for commercial deep-frying as this involves large volumes that are difficult to cool in a conventional fryer. This paper describes a method to minimize oil oxidation. French fries were deep-fried and the oil temperature decreased in a manner typical for a commercial deep-fryer. The concentration of polar compounds generated from thermally oxidized oil remarkably increased at temperature higher than 100°C but little oxidation occurred below 60°C. Heating the oil showed that the peroxide and polar compound content did not increase when the oil was actively cooled using a running water-cooled Graham-type condenser system to cool the oil from 180°C to room temperature within 30 min. When French fries were fried and the oil was then immediately cooled using the condenser, the polar compound content during cooling did not increase. Our results demonstrate that active cooling of heated oil is simple and quite effective for inhibiting oxidation.

  11. Radio-active waste disposal and deep-sea biology

    International Nuclear Information System (INIS)

    Rice, A.L.

    1978-01-01

    The deep-sea has been widely thought of as a remote, sparsely populated, and biologically inactive environment, well suited to receive the noxious products of nuclear fission processes. Much of what is known of abyssal biology tends to support this view, but there are a few disquieting contra-indications. The realisation, in recent years, that many animal groups show a previously unsuspected high species diversity in the deep-sea emphasized the paucity of our knowledge of this environment. More dramatically, the discovery of a large, active, and highly mobile abysso-bentho-pelagic fauna changed the whole concept of abyssal life. Finally, while there is little evidence for the existence of vertical migration patterns linking the deep-sea bottom communities with those of the overlying water layers, there are similarly too few negative results for the possibility of such transport mechanisms to be dismissed. In summary, biological knowledge of the abyss is insufficient to answer the questions raised in connection with deep-sea dumping, but in the absence of adequate answers it might be dangerous to ignore the questions

  12. Application of deep convolutional neural networks for ocean front recognition

    Science.gov (United States)

    Lima, Estanislau; Sun, Xin; Yang, Yuting; Dong, Junyu

    2017-10-01

    Ocean fronts have been a subject of study for many years, a variety of methods and algorithms have been proposed to address the problem of ocean fronts. However, all these existing ocean front recognition methods are built upon human expertise in defining the front based on subjective thresholds of relevant physical variables. This paper proposes a deep learning approach for ocean front recognition that is able to automatically recognize the front. We first investigated four existing deep architectures, i.e., AlexNet, CaffeNet, GoogLeNet, and VGGNet, for the ocean front recognition task using remote sensing (RS) data. We then propose a deep network with fewer layers compared to existing architecture for the front recognition task. This network has a total of five learnable layers. In addition, we extended the proposed network to recognize and classify the front into strong and weak ones. We evaluated and analyzed the proposed network with two strategies of exploiting the deep model: full-training and fine-tuning. Experiments are conducted on three different RS image datasets, which have different properties. Experimental results show that our model can produce accurate recognition results.

  13. Airline Passenger Profiling Based on Fuzzy Deep Machine Learning.

    Science.gov (United States)

    Zheng, Yu-Jun; Sheng, Wei-Guo; Sun, Xing-Ming; Chen, Sheng-Yong

    2017-12-01

    Passenger profiling plays a vital part of commercial aviation security, but classical methods become very inefficient in handling the rapidly increasing amounts of electronic records. This paper proposes a deep learning approach to passenger profiling. The center of our approach is a Pythagorean fuzzy deep Boltzmann machine (PFDBM), whose parameters are expressed by Pythagorean fuzzy numbers such that each neuron can learn how a feature affects the production of the correct output from both the positive and negative sides. We propose a hybrid algorithm combining a gradient-based method and an evolutionary algorithm for training the PFDBM. Based on the novel learning model, we develop a deep neural network (DNN) for classifying normal passengers and potential attackers, and further develop an integrated DNN for identifying group attackers whose individual features are insufficient to reveal the abnormality. Experiments on data sets from Air China show that our approach provides much higher learning ability and classification accuracy than existing profilers. It is expected that the fuzzy deep learning approach can be adapted for a variety of complex pattern analysis tasks.

  14. Deep Recurrent Neural Networks for Human Activity Recognition.

    Science.gov (United States)

    Murad, Abdulmajid; Pyun, Jae-Young

    2017-11-06

    Adopting deep learning methods for human activity recognition has been effective in extracting discriminative features from raw input sequences acquired from body-worn sensors. Although human movements are encoded in a sequence of successive samples in time, typical machine learning methods perform recognition tasks without exploiting the temporal correlations between input data samples. Convolutional neural networks (CNNs) address this issue by using convolutions across a one-dimensional temporal sequence to capture dependencies among input data. However, the size of convolutional kernels restricts the captured range of dependencies between data samples. As a result, typical models are unadaptable to a wide range of activity-recognition configurations and require fixed-length input windows. In this paper, we propose the use of deep recurrent neural networks (DRNNs) for building recognition models that are capable of capturing long-range dependencies in variable-length input sequences. We present unidirectional, bidirectional, and cascaded architectures based on long short-term memory (LSTM) DRNNs and evaluate their effectiveness on miscellaneous benchmark datasets. Experimental results show that our proposed models outperform methods employing conventional machine learning, such as support vector machine (SVM) and k-nearest neighbors (KNN). Additionally, the proposed models yield better performance than other deep learning techniques, such as deep believe networks (DBNs) and CNNs.

  15. Nuclear waste and a deep geological disposal facility

    International Nuclear Information System (INIS)

    Vokal, A.; Laciok, A.; Vasa, I.

    2005-01-01

    The paper presents a systematic analysis of the individual areas of research into nuclear waste and deep geological disposal with emphasis on the contribution of Nuclear Research Institute Rez plc to such efforts within international projects, specifically the EURATOM 6th Framework Programme. Research in the area of new advanced fuel cycles with focus on waste minimisation is based on EU's REDIMPACT project. The individual fuel cycles, which are currently studied within the EU, are briefly described. Special attention is paid to fast breeders and accelerator-driven reactor concepts associated with new spent fuel reprocessing technologies. Results obtained so far show that none even of the most advanced fuel cycles, currently under consideration, would eliminate the necessity to have a deep geological repository for a safe storage of residual radioactive waste. As regards deep geological repository barriers, the fact is highlighted that the safety of a repository is assured by complementary engineered and natural barriers. In order to demonstrate the safety of a repository, a deep insight must be gained into any and all of the individual processes that occur inside the repository and thus may affect radioactivity releases beyond the repository boundaries. The final section of the paper describes methods of radioactive waste conditioning for its disposal in a repository. Research into waste matrices used for radionuclide immobilisation is also highlighted. (author)

  16. Deep Neural Network-Based Chinese Semantic Role Labeling

    Institute of Scientific and Technical Information of China (English)

    ZHENG Xiaoqing; CHEN Jun; SHANG Guoqiang

    2017-01-01

    A recent trend in machine learning is to use deep architec-tures to discover multiple levels of features from data, which has achieved impressive results on various natural language processing (NLP) tasks. We propose a deep neural network-based solution to Chinese semantic role labeling (SRL) with its application on message analysis. The solution adopts a six-step strategy: text normalization, named entity recognition (NER), Chinese word segmentation and part-of-speech (POS) tagging, theme classification, SRL, and slot filling. For each step, a novel deep neural network - based model is designed and optimized, particularly for smart phone applications. Ex-periment results on all the NLP sub - tasks of the solution show that the proposed neural networks achieve state-of-the-art performance with the minimal computational cost. The speed advantage of deep neural networks makes them more competitive for large-scale applications or applications requir-ing real-time response, highlighting the potential of the pro-posed solution for practical NLP systems.

  17. Deep Borehole Field Test Requirements and Controlled Assumptions.

    Energy Technology Data Exchange (ETDEWEB)

    Hardin, Ernest [Sandia National Lab. (SNL-NM), Albuquerque, NM (United States)

    2015-07-01

    This document presents design requirements and controlled assumptions intended for use in the engineering development and testing of: 1) prototype packages for radioactive waste disposal in deep boreholes; 2) a waste package surface handling system; and 3) a subsurface system for emplacing and retrieving packages in deep boreholes. Engineering development and testing is being performed as part of the Deep Borehole Field Test (DBFT; SNL 2014a). This document presents parallel sets of requirements for a waste disposal system and for the DBFT, showing the close relationship. In addition to design, it will also inform planning for drilling, construction, and scientific characterization activities for the DBFT. The information presented here follows typical preparations for engineering design. It includes functional and operating requirements for handling and emplacement/retrieval equipment, waste package design and emplacement requirements, borehole construction requirements, sealing requirements, and performance criteria. Assumptions are included where they could impact engineering design. Design solutions are avoided in the requirements discussion. Deep Borehole Field Test Requirements and Controlled Assumptions July 21, 2015 iv ACKNOWLEDGEMENTS This set of requirements and assumptions has benefited greatly from reviews by Gordon Appel, Geoff Freeze, Kris Kuhlman, Bob MacKinnon, Steve Pye, David Sassani, Dave Sevougian, and Jiann Su.

  18. Equivalent drawbead performance in deep drawing simulations

    NARCIS (Netherlands)

    Meinders, Vincent T.; Geijselaers, Hubertus J.M.; Huetink, Han

    1999-01-01

    Drawbeads are applied in the deep drawing process to improve the control of the material flow during the forming operation. In simulations of the deep drawing process these drawbeads can be replaced by an equivalent drawbead model. In this paper the usage of an equivalent drawbead model in the

  19. Is deep dreaming the new collage?

    Science.gov (United States)

    Boden, Margaret A.

    2017-10-01

    Deep dreaming (DD) can combine and transform images in surprising ways. But, being based in deep learning (DL), it is not analytically understood. Collage is an art form that is constrained along various dimensions. DD will not be able to generate collages until DL can be guided in a disciplined fashion.

  20. Deep web search: an overview and roadmap

    NARCIS (Netherlands)

    Tjin-Kam-Jet, Kien; Trieschnigg, Rudolf Berend; Hiemstra, Djoerd

    2011-01-01

    We review the state-of-the-art in deep web search and propose a novel classification scheme to better compare deep web search systems. The current binary classification (surfacing versus virtual integration) hides a number of implicit decisions that must be made by a developer. We make these

  1. Research Proposal for Distributed Deep Web Search

    NARCIS (Netherlands)

    Tjin-Kam-Jet, Kien

    2010-01-01

    This proposal identifies two main problems related to deep web search, and proposes a step by step solution for each of them. The first problem is about searching deep web content by means of a simple free-text interface (with just one input field, instead of a complex interface with many input

  2. Development of Hydro-Mechanical Deep Drawing

    DEFF Research Database (Denmark)

    Zhang, Shi-Hong; Danckert, Joachim

    1998-01-01

    The hydro-mechanical deep-drawing process is reviewed in this article. The process principles and features are introduced and the developments of the hydro-mechanical deep-drawing process in process performances, in theory and in numerical simulation are described. The applications are summarized....... Some other related hydraulic forming processes are also dealt with as a comparison....

  3. Temperature impacts on deep-sea biodiversity.

    Science.gov (United States)

    Yasuhara, Moriaki; Danovaro, Roberto

    2016-05-01

    Temperature is considered to be a fundamental factor controlling biodiversity in marine ecosystems, but precisely what role temperature plays in modulating diversity is still not clear. The deep ocean, lacking light and in situ photosynthetic primary production, is an ideal model system to test the effects of temperature changes on biodiversity. Here we synthesize current knowledge on temperature-diversity relationships in the deep sea. Our results from both present and past deep-sea assemblages suggest that, when a wide range of deep-sea bottom-water temperatures is considered, a unimodal relationship exists between temperature and diversity (that may be right skewed). It is possible that temperature is important only when at relatively high and low levels but does not play a major role in the intermediate temperature range. Possible mechanisms explaining the temperature-biodiversity relationship include the physiological-tolerance hypothesis, the metabolic hypothesis, island biogeography theory, or some combination of these. The possible unimodal relationship discussed here may allow us to identify tipping points at which on-going global change and deep-water warming may increase or decrease deep-sea biodiversity. Predicted changes in deep-sea temperatures due to human-induced climate change may have more adverse consequences than expected considering the sensitivity of deep-sea ecosystems to temperature changes. © 2014 Cambridge Philosophical Society.

  4. A nonintrusive temperature measuring system for estimating deep body temperature in bed.

    Science.gov (United States)

    Sim, S Y; Lee, W K; Baek, H J; Park, K S

    2012-01-01

    Deep body temperature is an important indicator that reflects human being's overall physiological states. Existing deep body temperature monitoring systems are too invasive to apply to awake patients for a long time. Therefore, we proposed a nonintrusive deep body temperature measuring system. To estimate deep body temperature nonintrusively, a dual-heat-flux probe and double-sensor probes were embedded in a neck pillow. When a patient uses the neck pillow to rest, the deep body temperature can be assessed using one of the thermometer probes embedded in the neck pillow. We could estimate deep body temperature in 3 different sleep positions. Also, to reduce the initial response time of dual-heat-flux thermometer which measures body temperature in supine position, we employed the curve-fitting method to one subject. And thereby, we could obtain the deep body temperature in a minute. This result shows the possibility that the system can be used as practical temperature monitoring system with appropriate curve-fitting model. In the next study, we would try to establish a general fitting model that can be applied to all of the subjects. In addition, we are planning to extract meaningful health information such as sleep structure analysis from deep body temperature data which are acquired from this system.

  5. Deep-sea biodiversity in the Mediterranean Sea: the known, the unknown, and the unknowable.

    Directory of Open Access Journals (Sweden)

    Roberto Danovaro

    Full Text Available Deep-sea ecosystems represent the largest biome of the global biosphere, but knowledge of their biodiversity is still scant. The Mediterranean basin has been proposed as a hot spot of terrestrial and coastal marine biodiversity but has been supposed to be impoverished of deep-sea species richness. We summarized all available information on benthic biodiversity (Prokaryotes, Foraminifera, Meiofauna, Macrofauna, and Megafauna in different deep-sea ecosystems of the Mediterranean Sea (200 to more than 4,000 m depth, including open slopes, deep basins, canyons, cold seeps, seamounts, deep-water corals and deep-hypersaline anoxic basins and analyzed overall longitudinal and bathymetric patterns. We show that in contrast to what was expected from the sharp decrease in organic carbon fluxes and reduced faunal abundance, the deep-sea biodiversity of both the eastern and the western basins of the Mediterranean Sea is similarly high. All of the biodiversity components, except Bacteria and Archaea, displayed a decreasing pattern with increasing water depth, but to a different extent for each component. Unlike patterns observed for faunal abundance, highest negative values of the slopes of the biodiversity patterns were observed for Meiofauna, followed by Macrofauna and Megafauna. Comparison of the biodiversity associated with open slopes, deep basins, canyons, and deep-water corals showed that the deep basins were the least diverse. Rarefaction curves allowed us to estimate the expected number of species for each benthic component in different bathymetric ranges. A large fraction of exclusive species was associated with each specific habitat or ecosystem. Thus, each deep-sea ecosystem contributes significantly to overall biodiversity. From theoretical extrapolations we estimate that the overall deep-sea Mediterranean biodiversity (excluding prokaryotes reaches approximately 2805 species of which about 66% is still undiscovered. Among the biotic components

  6. Best in show but not best shape: a photographic assessment of show dog body condition.

    Science.gov (United States)

    Such, Z R; German, A J

    2015-08-01

    Previous studies suggest that owners often wrongly perceive overweight dogs to be in normal condition. The body shape of dogs attending shows might influence owners' perceptions, with online images of overweight show winners having a negative effect. This was an observational in silico study of canine body condition. 14 obese-prone breeds and 14 matched non-obese-probe breeds were first selected, and one operator then used an online search engine to identify 40 images, per breed, of dogs that had appeared at a major national UK show (Crufts). After images were anonymised and coded, a second observer subjectively assessed body condition, in a single sitting, using a previously validated method. Of 1120 photographs initially identified, 960 were suitable for assessing body condition, with all unsuitable images being from longhaired breeds. None of the dogs (0 per cent) were underweight, 708 (74 per cent) were in ideal condition and 252 (26 per cent) were overweight. Pugs, basset hounds and Labrador retrievers were most likely to be overweight, while standard poodles, Rhodesian ridgebacks, Hungarian vizslas and Dobermanns were least likely to be overweight. Given the proportion of show dogs from some breeds that are overweight, breed standards should be redefined to be consistent with a dog in optimal body condition. British Veterinary Association.

  7. Behavior of candidate canister materials in deep ocean environments

    International Nuclear Information System (INIS)

    Smyrl, W.H.; Stephenson, L.L.; Braithwaite, J.W.

    1977-04-01

    Corrosion tests have been conducted under simulated deep ocean conditions for nine months. The materials tested were base alloys of titanium, zirconium, and nickel. All materials tested showed corrosion rates that were very low even at the highest test temperature. None showed susceptibility to either stress corrosion cracking or differential aeration corrosion. Ambient electrochemical tests confirmed the findings that none should be sensitive to differential oxygen effects. The zirconium alloys may be more susceptible to pitting corrosion than the others, although the pitting conditions are unlikely to be found in service, unless higher temperatures are encountered. All the alloys tested could give long life under deep ocean conditions and are candidates for more detailed corrosion studies

  8. Deep subcritical levels measurements dependents upon kinetic distortion factors

    International Nuclear Information System (INIS)

    Pan Shibiao; Li Xiang; Fu Guo'en; Huang Liyuan; Mu Keliang

    2013-01-01

    The measurement of deep subcritical levels, with the increase of subcriticality, showed that the results impact on the kinetic distortion effect, along with neutron flux strongly deteriorated. Using the diffusion theory, calculations have been carried out to quantify the kinetic distortion correction factors in subcritical systems, and these indicate that epithermal neutron distributions are strongly affected by kinetic distortion. Subcriticality measurements in four different rod-state combination at the zero power device was carried out. The test data analysis shows that, with increasing subcriticality, kinetic distortion effect correction factor gradually increases from 1.052 to 1.065, corresponding reactive correction amount of 0.78β eff ∼ 3.01β eff . Thus, it is necessary to consider the kinetic distortion effect in the deep subcritical reactivity measurements. (authors)

  9. Deep-inelastic scattering in 124,136Xe+58,64Ni at energies near the Coulomb barrier

    International Nuclear Information System (INIS)

    Gehring, J.; Back, B.B.; Chan, K.C.; Freer, M.; Henderson, D.; Jiang, C.L.; Rehm, K.E.; Schiffer, J.P.; Wolanski, M.; Wuosmaa, A.H.; Gehring, J.; Wolanski, M.

    1997-01-01

    Cross sections, angular distributions, and mass distributions have been measured for deep-inelastic scattering in 124 Xe+ 58 Ni and 136 Xe+ 64 Ni at laboratory energies in the vicinity of the Coulomb barrier. The mass distributions show distinct components due to deep-inelastic and fissionlike processes. The strength of deep-inelastic scattering is similar in the two systems measured and comparable to previous measurements in 58 Ni+ 112,124 Sn. copyright 1997 The American Physical Society

  10. Morphology of germlings of urediniospores and its value for the identification and classification of grass rust fungi

    NARCIS (Netherlands)

    Swertz, C.A.

    1994-01-01

    The identification and classification of grass rust fungi is often difficult since most traditionally used morphological characters are quantitative and subjective. Besides, when using the host range as a taxonomic criterion, it is important to realize that a rust fungus may have jumped to

  11. Analyses of the deep borehole drilling status for a deep borehole disposal system

    Energy Technology Data Exchange (ETDEWEB)

    Lee, Jong Youl; Choi, Heui Joo; Lee, Min Soo; Kim, Geon Young; Kim, Kyung Su [KAERI, Daejeon (Korea, Republic of)

    2016-05-15

    The purpose of disposal for radioactive wastes is not only to isolate them from humans, but also to inhibit leakage of any radioactive materials into the accessible environment. Because of the extremely high level and long-time scale radioactivity of HLW(High-level radioactive waste), a mined deep geological disposal concept, the disposal depth is about 500 m below ground, is considered as the safest method to isolate the spent fuels or high-level radioactive waste from the human environment with the best available technology at present time. Therefore, as an alternative disposal concept, i.e., deep borehole disposal technology is under consideration in number of countries in terms of its outstanding safety and cost effectiveness. In this paper, the general status of deep drilling technologies was reviewed for deep borehole disposal of high level radioactive wastes. Based on the results of these review, very preliminary applicability of deep drilling technology for deep borehole disposal analyzed. In this paper, as one of key technologies of deep borehole disposal system, the general status of deep drilling technologies in oil industry, geothermal industry and geo scientific field was reviewed for deep borehole disposal of high level radioactive wastes. Based on the results of these review, the very preliminary applicability of deep drilling technology for deep borehole disposal such as relation between depth and diameter, drilling time and feasibility classification was analyzed.

  12. The Implications of Deep Mitigation Pathways

    Science.gov (United States)

    Calvin, K. V.

    2016-12-01

    The 21st Conference of Parties to the UNFCCC agreement called for limiting climate change to "well below 2°C above pre-industrial levels and pursuing efforts to limit the temperature increase to 1.5°C." A climate target of 1.5°C places a stringent constraint on allowable emissions over the twenty-first century. Roegli et al. (2015) set that constraint at 200-415 GtCO2 between 2011 and 2100 for a likely chance of staying below 1.5°C in 2100. Limiting emissions to these levels requires that global emissions peak and decline over the coming decades, with net negative global emissions by mid-century. This level of decarbonization requires dramatic shifts in the energy and agricultural sectors, and comes at significant economic costs. This talk explores the effect of mitigating climate change to 1.5°C on the economy, energy system, and terrestrial system. We quantify the required deployment of various low carbon technologies, as well as the amount of existing capital that is abandoned in an effort to limit emissions. We show the shifts required in the terrestrial system, including its contribution to carbon sequestration through afforestation and bioenergy. Additionally, we show the implications of deep mitigation pathways on energy, food, and carbon prices. We contrast these results with a reference, no climate policy, world and a 2°C.

  13. Hot-carrier effects on irradiated deep submicron NMOSFET

    International Nuclear Information System (INIS)

    Cui Jiangwei; Zheng Qiwen; Yu Xuefeng; Cong Zhongchao; Zhou Hang; Guo Qi; Wen Lin; Wei Ying; Ren Diyuan

    2014-01-01

    We investigate how γ exposure impacts the hot-carrier degradation in deep submicron NMOSFET with different technologies and device geometries for the first time. The results show that hot-carrier degradations on irradiated devices are greater than those without irradiation, especially for narrow channel device. The reason is attributed to charge traps in STI, which then induce different electric field and impact ionization rates during hot-carrier stress. (semiconductor devices)

  14. A reason of fast and deep fading of centimeter wave

    International Nuclear Information System (INIS)

    Danzan, D.; Damdinsuren, E.; Hiamjav, J.; Chuluunbaatar, Ch.; Battulga, S.

    1992-01-01

    First discovered experimentally exactly correlation between of appearance and of disappearance of optical mirage and fast and deep fading of horizontal polarization of centimeter wave. Proved the interference of the straight and reflected rays from the thin layer of air in mirage a reason of this fading. The physical parameters data of the layer of mirage: change of dielectric permeability and n/ h gradient of refraction index of air in this layer are been showed

  15. Microbial investigations of deep geological compartments

    International Nuclear Information System (INIS)

    Barsotti, V.; Sergeant, C.; Vesvres, M.H.; Joulian, C.; Coulon, S.; Le Marrec, C.; Garrido, F.

    2010-01-01

    Document available in extended abstract form only. Deep sedimentary rocks are now considered to contain a significant part of the total bacterial population, but are microbiologically unexplored. The drilling down to the base of the Triassic (1980 meters deep) in the geological formations of the eastern Paris Basin performed by ANDRA (EST433) in 2008 provides us a good opportunity to explore the deep biosphere. We conditioned and sub-sampled on the coring site, in as aseptic conditions as possible, the nine cores: two in the Callovo-Oxfordian clay, two in the Dogger, five in the Triassic compartments. In addition to storage at atmospheric pressure, a portion of the five Triassic samples was placed in a 190 bars pressurized bars chamber to investigate the influence of the conservation pressure factor on the found microflora. In parallel, in order to evaluate a potential bacterial contamination of the core by the drilling fluids, samples of mud just before each sample drilling were taken and analysed. The microbial exploration we started can be divided in two parts: - A cultural approach in different culture media for six metabolic groups to try to find microbial cells still viable. This type of experiment is difficult because of the small proportion of cultivable species, especially in these extreme environmental samples. - A molecular approach by direct extraction of genomic DNA from the geological samples to explore a larger biodiversity. Here, the limits are the difficulties to extract DNA from these low biomass containing rocks. The five Triassic samples were partly crushed in powder and inoculated in the six culture media with four NaCl concentrations, because this type of rock is known as saline or hyper-saline, and incubated at three temperatures: 30 deg. C, 55 deg. C under agitation and 70 deg. C. First results will be presented. The direct extraction of DNA needs a complete method optimisation to adapt existent procedures (using commercial kit and classical

  16. The deep, hot biosphere: Twenty-five years of retrospection.

    Science.gov (United States)

    Colman, Daniel R; Poudel, Saroj; Stamps, Blake W; Boyd, Eric S; Spear, John R

    2017-07-03

    Twenty-five years ago this month, Thomas Gold published a seminal manuscript suggesting the presence of a "deep, hot biosphere" in the Earth's crust. Since this publication, a considerable amount of attention has been given to the study of deep biospheres, their role in geochemical cycles, and their potential to inform on the origin of life and its potential outside of Earth. Overwhelming evidence now supports the presence of a deep biosphere ubiquitously distributed on Earth in both terrestrial and marine settings. Furthermore, it has become apparent that much of this life is dependent on lithogenically sourced high-energy compounds to sustain productivity. A vast diversity of uncultivated microorganisms has been detected in subsurface environments, and we show that H 2 , CH 4 , and CO feature prominently in many of their predicted metabolisms. Despite 25 years of intense study, key questions remain on life in the deep subsurface, including whether it is endemic and the extent of its involvement in the anaerobic formation and degradation of hydrocarbons. Emergent data from cultivation and next-generation sequencing approaches continue to provide promising new hints to answer these questions. As Gold suggested, and as has become increasingly evident, to better understand the subsurface is critical to further understanding the Earth, life, the evolution of life, and the potential for life elsewhere. To this end, we suggest the need to develop a robust network of interdisciplinary scientists and accessible field sites for long-term monitoring of the Earth's subsurface in the form of a deep subsurface microbiome initiative.

  17. Deep transfer learning for automatic target classification: MWIR to LWIR

    Science.gov (United States)

    Ding, Zhengming; Nasrabadi, Nasser; Fu, Yun

    2016-05-01

    Publisher's Note: This paper, originally published on 5/12/2016, was replaced with a corrected/revised version on 5/18/2016. If you downloaded the original PDF but are unable to access the revision, please contact SPIE Digital Library Customer Service for assistance. When dealing with sparse or no labeled data in the target domain, transfer learning shows its appealing performance by borrowing the supervised knowledge from external domains. Recently deep structure learning has been exploited in transfer learning due to its attractive power in extracting effective knowledge through multi-layer strategy, so that deep transfer learning is promising to address the cross-domain mismatch. In general, cross-domain disparity can be resulted from the difference between source and target distributions or different modalities, e.g., Midwave IR (MWIR) and Longwave IR (LWIR). In this paper, we propose a Weighted Deep Transfer Learning framework for automatic target classification through a task-driven fashion. Specifically, deep features and classifier parameters are obtained simultaneously for optimal classification performance. In this way, the proposed deep structures can extract more effective features with the guidance of the classifier performance; on the other hand, the classifier performance is further improved since it is optimized on more discriminative features. Furthermore, we build a weighted scheme to couple source and target output by assigning pseudo labels to target data, therefore we can transfer knowledge from source (i.e., MWIR) to target (i.e., LWIR). Experimental results on real databases demonstrate the superiority of the proposed algorithm by comparing with others.

  18. Deep-Sea, Deep-Sequencing: Metabarcoding Extracellular DNA from Sediments of Marine Canyons.

    Directory of Open Access Journals (Sweden)

    Magdalena Guardiola

    Full Text Available Marine sediments are home to one of the richest species pools on Earth, but logistics and a dearth of taxonomic work-force hinders the knowledge of their biodiversity. We characterized α- and β-diversity of deep-sea assemblages from submarine canyons in the western Mediterranean using an environmental DNA metabarcoding. We used a new primer set targeting a short eukaryotic 18S sequence (ca. 110 bp. We applied a protocol designed to obtain extractions enriched in extracellular DNA from replicated sediment corers. With this strategy we captured information from DNA (local or deposited from the water column that persists adsorbed to inorganic particles and buffered short-term spatial and temporal heterogeneity. We analysed replicated samples from 20 localities including 2 deep-sea canyons, 1 shallower canal, and two open slopes (depth range 100-2,250 m. We identified 1,629 MOTUs, among which the dominant groups were Metazoa (with representatives of 19 phyla, Alveolata, Stramenopiles, and Rhizaria. There was a marked small-scale heterogeneity as shown by differences in replicates within corers and within localities. The spatial variability between canyons was significant, as was the depth component in one of the canyons where it was tested. Likewise, the composition of the first layer (1 cm of sediment was significantly different from deeper layers. We found that qualitative (presence-absence and quantitative (relative number of reads data showed consistent trends of differentiation between samples and geographic areas. The subset of exclusively benthic MOTUs showed similar patterns of β-diversity and community structure as the whole dataset. Separate analyses of the main metazoan phyla (in number of MOTUs showed some differences in distribution attributable to different lifestyles. Our results highlight the differentiation that can be found even between geographically close assemblages, and sets the ground for future monitoring and conservation

  19. Impact disruption and recovery of the deep subsurface biosphere

    Science.gov (United States)

    Cockell, Charles S.; Voytek, Mary A.; Gronstal, Aaron L.; Finster, Kai; Kirshtein, Julie D.; Howard, Kieren; Reitner, Joachim; Gohn, Gregory S.; Sanford, Ward E.; Horton, J. Wright; Kallmeyer, Jens; Kelly, Laura; Powars, David S.

    2012-01-01

    Although a large fraction of the world's biomass resides in the subsurface, there has been no study of the effects of catastrophic disturbance on the deep biosphere and the rate of its subsequent recovery. We carried out an investigation of the microbiology of a 1.76 km drill core obtained from the ~35 million-year-old Chesapeake Bay impact structure, USA, with robust contamination control. Microbial enumerations displayed a logarithmic downward decline, but the different gradient, when compared to previously studied sites, and the scatter of the data are consistent with a microbiota influenced by the geological disturbances caused by the impact. Microbial abundance is low in buried crater-fill, ocean-resurge, and avalanche deposits despite the presence of redox couples for growth. Coupled with the low hydraulic conductivity, the data suggest the microbial community has not yet recovered from the impact ~35 million years ago. Microbial enumerations, molecular analysis of microbial enrichment cultures, and geochemical analysis showed recolonization of a deep region of impact-fractured rock that was heated to above the upper temperature limit for life at the time of impact. These results show how, by fracturing subsurface rocks, impacts can extend the depth of the biosphere. This phenomenon would have provided deep refugia for life on the more heavily bombarded early Earth, and it shows that the deeply fractured regions of impact craters are promising targets to study the past and present habitability of Mars.

  20. WFIRST: Science from Deep Field Surveys

    Science.gov (United States)

    Koekemoer, Anton; Foley, Ryan; WFIRST Deep Field Working Group

    2018-01-01

    WFIRST will enable deep field imaging across much larger areas than those previously obtained with Hubble, opening up completely new areas of parameter space for extragalactic deep fields including cosmology, supernova and galaxy evolution science. The instantaneous field of view of the Wide Field Instrument (WFI) is about 0.3 square degrees, which would for example yield an Ultra Deep Field (UDF) reaching similar depths at visible and near-infrared wavelengths to that obtained with Hubble, over an area about 100-200 times larger, for a comparable investment in time. Moreover, wider fields on scales of 10-20 square degrees could achieve depths comparable to large HST surveys at medium depths such as GOODS and CANDELS, and would enable multi-epoch supernova science that could be matched in area to LSST Deep Drilling fields or other large survey areas. Such fields may benefit from being placed on locations in the sky that have ancillary multi-band imaging or spectroscopy from other facilities, from the ground or in space. The WFIRST Deep Fields Working Group has been examining the science considerations for various types of deep fields that may be obtained with WFIRST, and present here a summary of the various properties of different locations in the sky that may be considered for future deep fields with WFIRST.

  1. The deep lymphatic anatomy of the hand.

    Science.gov (United States)

    Ma, Chuan-Xiang; Pan, Wei-Ren; Liu, Zhi-An; Zeng, Fan-Qiang; Qiu, Zhi-Qiang

    2018-04-03

    The deep lymphatic anatomy of the hand still remains the least described in medical literature. Eight hands were harvested from four nonembalmed human cadavers amputated above the wrist. A small amount of 6% hydrogen peroxide was employed to detect the lymphatic vessels around the superficial and deep palmar vascular arches, in webs from the index to little fingers, the thenar and hypothenar areas. A 30-gauge needle was inserted into the vessels and injected with a barium sulphate compound. Each specimen was dissected, photographed and radiographed to demonstrate deep lymphatic distribution of the hand. Five groups of deep collecting lymph vessels were found in the hand: superficial palmar arch lymph vessel (SPALV); deep palmar arch lymph vessel (DPALV); thenar lymph vessel (TLV); hypothenar lymph vessel (HTLV); deep finger web lymph vessel (DFWLV). Each group of vessels drained in different directions first, then all turned and ran towards the wrist in different layers. The deep lymphatic drainage of the hand has been presented. The results will provide an anatomical basis for clinical management, educational reference and scientific research. Copyright © 2018 Elsevier GmbH. All rights reserved.

  2. Tomato Fruits Show Wide Phenomic Diversity but Fruit Developmental Genes Show Low Genomic Diversity.

    Directory of Open Access Journals (Sweden)

    Vijee Mohan

    Full Text Available Domestication of tomato has resulted in large diversity in fruit phenotypes. An intensive phenotyping of 127 tomato accessions from 20 countries revealed extensive morphological diversity in fruit traits. The diversity in fruit traits clustered the accessions into nine classes and identified certain promising lines having desirable traits pertaining to total soluble salts (TSS, carotenoids, ripening index, weight and shape. Factor analysis of the morphometric data from Tomato Analyzer showed that the fruit shape is a complex trait shared by several factors. The 100% variance between round and flat fruit shapes was explained by one discriminant function having a canonical correlation of 0.874 by stepwise discriminant analysis. A set of 10 genes (ACS2, COP1, CYC-B, RIN, MSH2, NAC-NOR, PHOT1, PHYA, PHYB and PSY1 involved in various plant developmental processes were screened for SNP polymorphism by EcoTILLING. The genetic diversity in these genes revealed a total of 36 non-synonymous and 18 synonymous changes leading to the identification of 28 haplotypes. The average frequency of polymorphism across the genes was 0.038/Kb. Significant negative Tajima'D statistic in two of the genes, ACS2 and PHOT1 indicated the presence of rare alleles in low frequency. Our study indicates that while there is low polymorphic diversity in the genes regulating plant development, the population shows wider phenotype diversity. Nonetheless, morphological and genetic diversity of the present collection can be further exploited as potential resources in future.

  3. Shear Strengthening of RC Deep Beam Using Externally Bonded GFRP Fabrics

    Science.gov (United States)

    Kumari, A.; Patel, S. S.; Nayak, A. N.

    2018-06-01

    This work presents the experimental investigation of RC deep beams wrapped with externally bonded Glass Fibre Reinforced Polymer (GFRP) fabrics in order to study the Load versus deflection behavior, cracking pattern, failure modes and ultimate shear strength. A total number of five deep beams have been casted, which is designed with conventional steel reinforcement as per IS: 456 (Indian standard plain and reinforced concrete—code for practice, Bureau of Indian Standards, New Delhi, 2000). The spans to depth ratio for all RC deep beams have been kept less than 2 as per the above specification. Out of five RC deep beams, one without retrofitting serves as a reference beam and the rest four have been wrapped with GFRP fabrics in multiple layers and tested with two point loading condition. The first cracking load, ultimate load and the shear contribution of GFRP to the deep beams have been observed. A critical discussion is made with respect to the enhancement of the strength, behaviour and performance of retrofitted deep beams in comparison to the deep beam without GFRP in order to explore the potential use of GFRP for strengthening the RC deep beams. Test results have demonstrated that the deep beams retrofitted with GFRP shows a slower development of the diagonal cracks and improves shear carrying capacity of the RC deep beam. A comparative study of the experimental results with the theoretical ones predicted by various researchers available in the literatures has also been presented. It is observed that the ultimate load of the beams retrofitted with GFRP fabrics increases with increase of number of GFRP layers up to a specific number of layers, i.e. 3 layers, beyond which it decreases.

  4. Sensory processing of deep tissue nociception in the rat spinal cord and thalamic ventrobasal complex.

    Science.gov (United States)

    Sikandar, Shafaq; West, Steven J; McMahon, Stephen B; Bennett, David L; Dickenson, Anthony H

    2017-07-01

    Sensory processing of deep somatic tissue constitutes an important component of the nociceptive system, yet associated central processing pathways remain poorly understood. Here, we provide a novel electrophysiological characterization and immunohistochemical analysis of neural activation in the lateral spinal nucleus (LSN). These neurons show evoked activity to deep, but not cutaneous, stimulation. The evoked responses of neurons in the LSN can be sensitized to somatosensory stimulation following intramuscular hypertonic saline, an acute model of muscle pain, suggesting this is an important spinal relay site for the processing of deep tissue nociceptive inputs. Neurons of the thalamic ventrobasal complex (VBC) mediate both cutaneous and deep tissue sensory processing, but in contrast to the lateral spinal nucleus our electrophysiological studies do not suggest the existence of a subgroup of cells that selectively process deep tissue inputs. The sensitization of polymodal and thermospecific VBC neurons to mechanical somatosensory stimulation following acute muscle stimulation with hypertonic saline suggests differential roles of thalamic subpopulations in mediating cutaneous and deep tissue nociception in pathological states. Overall, our studies at both the spinal (lateral spinal nucleus) and supraspinal (thalamic ventrobasal complex) levels suggest a convergence of cutaneous and deep somatosensory inputs onto spinothalamic pathways, which are unmasked by activation of muscle nociceptive afferents to produce consequent phenotypic alterations in spinal and thalamic neural coding of somatosensory stimulation. A better understanding of the sensory pathways involved in deep tissue nociception, as well as the degree of labeled line and convergent pathways for cutaneous and deep somatosensory inputs, is fundamental to developing targeted analgesic therapies for deep pain syndromes. © 2017 University College London. Physiological Reports published by Wiley Periodicals

  5. Inferences of the deep solar meridional flow

    Science.gov (United States)

    Böning, Vincent G. A.

    2017-10-01

    Understanding the solar meridional flow is important for uncovering the origin of the solar activity cycle. Yet, recent helioseismic estimates of this flow have come to conflicting conclusions in deeper layers of the solar interior, i.e., at depths below about 0.9 solar radii. The aim of this thesis is to contribute to a better understanding of the deep solar meridional flow. Time-distance helioseismology is the major method for investigating this flow. In this method, travel times of waves propagating between pairs of locations on the solar surface are measured. Until now, the travel-time measurements have been modeled using the ray approximation, which assumes that waves travel along infinitely thin ray paths between these locations. In contrast, the scattering of the full wave field in the solar interior due to the flow is modeled in first order by the Born approximation. It is in general a more accurate model of the physics in the solar interior. In a first step, an existing model for calculating the sensitivity of travel-time measurements to solar interior flows using the Born approximation is extended from Cartesian to spherical geometry. The results are succesfully compared to the Cartesian ones and are tested for self-consistency. In a second step, the newly developed model is validated using an existing numerical simulation of linear wave propagation in the Sun. An inversion of artificial travel times for meridional flow shows excellent agreement for noiseless data and reproduces many features in the input flow profile in the case of noisy data. Finally, the new method is used to infer the deep meridional flow. I used Global Oscillation Network Group (GONG) data that were earlier analyzed using the ray approximation and I employed the same Substractive Optimized Local Averaging (SOLA) inversion technique as in the earlier study. Using an existing formula for the covariance of travel-time measurements, it is shown that the assumption of uncorrelated errors

  6. Increasing the Deep Drawability of Al-1050 Aluminum Sheet using Multi-Point Blank Holder

    Directory of Open Access Journals (Sweden)

    Gavas, M.

    2006-01-01

    Full Text Available Aluminum alloys have been widely used in the fields of automobile and aerospace industries. Due to their bad cold-formability in deep drawing, a lot of forming methods have been implemented to increase the drawing height and the limiting drawing rate (LDR. The conventional deep drawing process is limited to a certain limit drawing ratio beyond which failure will ensue. The purpose of this experimental study is to examine the possibilities of increasing this limitation using the multi-point blank holder. The results from the experiments showed that the multi-point blank holder is effective way to promote deep drawability of Al-1050 sheet.

  7. Deep learning improves prediction of CRISPR-Cpf1 guide RNA activity.

    Science.gov (United States)

    Kim, Hui Kwon; Min, Seonwoo; Song, Myungjae; Jung, Soobin; Choi, Jae Woo; Kim, Younggwang; Lee, Sangeun; Yoon, Sungroh; Kim, Hyongbum Henry

    2018-03-01

    We present two algorithms to predict the activity of AsCpf1 guide RNAs. Indel frequencies for 15,000 target sequences were used in a deep-learning framework based on a convolutional neural network to train Seq-deepCpf1. We then incorporated chromatin accessibility information to create the better-performing DeepCpf1 algorithm for cell lines for which such information is available and show that both algorithms outperform previous machine learning algorithms on our own and published data sets.

  8. An improved advertising CTR prediction approach based on the fuzzy deep neural network.

    Science.gov (United States)

    Jiang, Zilong; Gao, Shu; Li, Mingjiang

    2018-01-01

    Combining a deep neural network with fuzzy theory, this paper proposes an advertising click-through rate (CTR) prediction approach based on a fuzzy deep neural network (FDNN). In this approach, fuzzy Gaussian-Bernoulli restricted Boltzmann machine (FGBRBM) is first applied to input raw data from advertising datasets. Next, fuzzy restricted Boltzmann machine (FRBM) is used to construct the fuzzy deep belief network (FDBN) with the unsupervised method layer by layer. Finally, fuzzy logistic regression (FLR) is utilized for modeling the CTR. The experimental results show that the proposed FDNN model outperforms several baseline models in terms of both data representation capability and robustness in advertising click log datasets with noise.

  9. The deep structure of the Sichuan basin and adjacent orogenic zones revealed by the aggregated deep seismic profiling datum

    Science.gov (United States)

    Xiong, X.; Gao, R.; Li, Q.; Wang, H.

    2012-12-01

    The sedimentary basin and the orogenic belt are the basic two tectonic units of the continental lithosphere, and form the basin-mountain coupling system, The research of which is the key element to the oil and gas exploration, the global tectonic theory and models and the development of the geological theory. The Sichuan basin and adjacent orogenic belts is one of the most ideal sites to research the issues above, in particular by the recent deep seismic profiling datum. From the 1980s to now, there are 11 deep seismic sounding profiles and 6 deep seismic reflection profiles and massive seismic broadband observation stations deployed around and crossed the Sichuan basin, which provide us a big opportunity to research the deep structure and other forward issues in this region. Supported by the National Natural Science Foundation of China (Grant No. 41104056) and the Fundamental Research Funds of the Institute of Geological Sciences, CAGS (No. J1119), we sampled the Moho depth and low-velocity zone depth and the Pn velocity of these datum, then formed the contour map of the Moho depth and Pn velocity by the interpolation of the sampled datum. The result shows the Moho depth beneath Sichuan basin ranges from 40 to 44 km, the sharp Moho offset appears in the western margin of the Sichuan basin, and there is a subtle Moho depression in the central southern part of the Sichuan basin; the P wave velocity can be 6.0 km/s at ca. 10 km deep, and increases gradually deeper, the average P wave velocity in this region is ca. 6.3 km/s; the Pn velocity is ca. 8.0-8.02 km/s in Sichuan basin, and 7.70-7.76 km/s in Chuan-Dian region; the low velocity zone appears in the western margin of the Sichuan basin, which maybe cause the cause of the earthquake.

  10. Hello World Deep Learning in Medical Imaging.

    Science.gov (United States)

    Lakhani, Paras; Gray, Daniel L; Pett, Carl R; Nagy, Paul; Shih, George

    2018-05-03

    There is recent popularity in applying machine learning to medical imaging, notably deep learning, which has achieved state-of-the-art performance in image analysis and processing. The rapid adoption of deep learning may be attributed to the availability of machine learning frameworks and libraries to simplify their use. In this tutorial, we provide a high-level overview of how to build a deep neural network for medical image classification, and provide code that can help those new to the field begin their informatics projects.

  11. Deep Generative Models for Molecular Science

    DEFF Research Database (Denmark)

    Jørgensen, Peter Bjørn; Schmidt, Mikkel Nørgaard; Winther, Ole

    2018-01-01

    Generative deep machine learning models now rival traditional quantum-mechanical computations in predicting properties of new structures, and they come with a significantly lower computational cost, opening new avenues in computational molecular science. In the last few years, a variety of deep...... generative models have been proposed for modeling molecules, which differ in both their model structure and choice of input features. We review these recent advances within deep generative models for predicting molecular properties, with particular focus on models based on the probabilistic autoencoder (or...

  12. Harnessing the Deep Web: Present and Future

    OpenAIRE

    Madhavan, Jayant; Afanasiev, Loredana; Antova, Lyublena; Halevy, Alon

    2009-01-01

    Over the past few years, we have built a system that has exposed large volumes of Deep-Web content to Google.com users. The content that our system exposes contributes to more than 1000 search queries per-second and spans over 50 languages and hundreds of domains. The Deep Web has long been acknowledged to be a major source of structured data on the web, and hence accessing Deep-Web content has long been a problem of interest in the data management community. In this paper, we report on where...

  13. Desalination Economic Evaluation Program (DEEP). User's manual

    International Nuclear Information System (INIS)

    2000-01-01

    DEEP (formerly named ''Co-generation and Desalination Economic Evaluation'' Spreadsheet, CDEE) has been developed originally by General Atomics under contract, and has been used in the IAEA's feasibility studies. For further confidence in the software, it was validated in March 1998. After that, a user friendly version has been issued under the name of DEEP at the end of 1998. DEEP output includes the levelised cost of water and power, a breakdown of cost components, energy consumption and net saleable power for each selected option. Specific power plants can be modelled by adjustment of input data including design power, power cycle parameters and costs

  14. Zooplankton at deep Red Sea brine pools

    KAUST Repository

    Kaartvedt, Stein

    2016-03-02

    The deep-sea anoxic brines of the Red Sea comprise unique, complex and extreme habitats. These environments are too harsh for metazoans, while the brine–seawater interface harbors dense microbial populations. We investigated the adjacent pelagic fauna at two brine pools using net tows, video records from a remotely operated vehicle and submerged echosounders. Waters just above the brine pool of Atlantis II Deep (2000 m depth) appeared depleted of macrofauna. In contrast, the fauna appeared to be enriched at the Kebrit Deep brine–seawater interface (1466 m).

  15. NATURAL GAS RESOURCES IN DEEP SEDIMENTARY BASINS

    Energy Technology Data Exchange (ETDEWEB)

    Thaddeus S. Dyman; Troy Cook; Robert A. Crovelli; Allison A. Henry; Timothy C. Hester; Ronald C. Johnson; Michael D. Lewan; Vito F. Nuccio; James W. Schmoker; Dennis B. Riggin; Christopher J. Schenk

    2002-02-05

    From a geological perspective, deep natural gas resources are generally defined as resources occurring in reservoirs at or below 15,000 feet, whereas ultra-deep gas occurs below 25,000 feet. From an operational point of view, ''deep'' is often thought of in a relative sense based on the geologic and engineering knowledge of gas (and oil) resources in a particular area. Deep gas can be found in either conventionally-trapped or unconventional basin-center accumulations that are essentially large single fields having spatial dimensions often exceeding those of conventional fields. Exploration for deep conventional and unconventional basin-center natural gas resources deserves special attention because these resources are widespread and occur in diverse geologic environments. In 1995, the U.S. Geological Survey estimated that 939 TCF of technically recoverable natural gas remained to be discovered or was part of reserve appreciation from known fields in the onshore areas and State waters of the United. Of this USGS resource, nearly 114 trillion cubic feet (Tcf) of technically-recoverable gas remains to be discovered from deep sedimentary basins. Worldwide estimates of deep gas are also high. The U.S. Geological Survey World Petroleum Assessment 2000 Project recently estimated a world mean undiscovered conventional gas resource outside the U.S. of 844 Tcf below 4.5 km (about 15,000 feet). Less is known about the origins of deep gas than about the origins of gas at shallower depths because fewer wells have been drilled into the deeper portions of many basins. Some of the many factors contributing to the origin of deep gas include the thermal stability of methane, the role of water and non-hydrocarbon gases in natural gas generation, porosity loss with increasing thermal maturity, the kinetics of deep gas generation, thermal cracking of oil to gas, and source rock potential based on thermal maturity and kerogen type. Recent experimental simulations

  16. Potential Osteoporosis Recovery by Deep Sea Water through Bone Regeneration in SAMP8 Mice

    Directory of Open Access Journals (Sweden)

    Hen-Yu Liu

    2013-01-01

    Full Text Available The aim of this study is to examine the therapeutic potential of deep sea water (DSW on osteoporosis. Previously, we have established the ovariectomized senescence-accelerated mice (OVX-SAMP8 and demonstrated strong recovery of osteoporosis by stem cell and platelet-rich plasma (PRP. Deep sea water at hardness (HD 1000 showed significant increase in proliferation of osteoblastic cell (MC3T3 by MTT assay. For in vivo animal study, bone mineral density (BMD was strongly enhanced followed by the significantly increased trabecular numbers through micro-CT examination after a 4-month deep sea water treatment, and biochemistry analysis showed that serum alkaline phosphatase (ALP activity was decreased. For stage-specific osteogenesis, bone marrow-derived stromal cells (BMSCs were harvested and examined. Deep sea water-treated BMSCs showed stronger osteogenic differentiation such as BMP2, RUNX2, OPN, and OCN, and enhanced colony forming abilities, compared to the control group. Interestingly, most untreated OVX-SAMP8 mice died around 10 months; however, approximately 57% of DSW-treated groups lived up to 16.6 months, a life expectancy similar to the previously reported life expectancy for SAMR1 24 months. The results demonstrated the regenerative potentials of deep sea water on osteogenesis, showing that deep sea water could potentially be applied in osteoporosis therapy as a complementary and alternative medicine (CAM.

  17. Connectivity between surface and deep waters determines prokaryotic diversity in the North Atlantic Deep Water.

    Science.gov (United States)

    Frank, Alexander H; Garcia, Juan A L; Herndl, Gerhard J; Reinthaler, Thomas

    2016-06-01

    To decipher the influence of depth stratification and surface provincialism on the dark ocean prokaryotic community composition, we sampled the major deep-water masses in the eastern North Atlantic covering three biogeographic provinces. Their diversity was evaluated using ordination and canonical analysis of 454 pyrotag sequences. Variance partitioning suggested that 16% of the variation in the bacterial community composition was based on depth stratification while 9% of the variation was due to geographic location. General linear mixed effect models showed that the community of the subsurface waters was connected to the dark ocean prokaryotic communities in different biogeographic provinces. Cluster analysis indicated that some prokaryotic taxa are specific to distinct regions in bathypelagic water masses. Taken together, our data suggest that the dark ocean prokaryotic community composition of the eastern North Atlantic is primed by the formation and the horizontal transport of water masses. © 2016 Society for Applied Microbiology and John Wiley & Sons Ltd.

  18. Comet Dust After Deep Impact

    Science.gov (United States)

    Wooden, Diane H.; Harker, David E.; Woodward, Charles E.

    2006-01-01

    When the Deep Impact Mission hit Jupiter Family comet 9P/Tempel 1, an ejecta crater was formed and an pocket of volatile gases and ices from 10-30 m below the surface was exposed (A Hearn et aI. 2005). This resulted in a gas geyser that persisted for a few hours (Sugita et al, 2005). The gas geyser pushed dust grains into the coma (Sugita et a1. 2005), as well as ice grains (Schulz et al. 2006). The smaller of the dust grains were submicron in radii (0-25.3 micron), and were primarily composed of highly refractory minerals including amorphous (non-graphitic) carbon, and silicate minerals including amorphous (disordered) olivine (Fe,Mg)2SiO4 and pyroxene (Fe,Mg)SiO3 and crystalline Mg-rich olivine. The smaller grains moved faster, as expected from the size-dependent velocity law produced by gas-drag on grains. The mineralogy evolved with time: progressively larger grains persisted in the near nuclear region, having been imparted with slower velocities, and the mineralogies of these larger grains appeared simpler and without crystals. The smaller 0.2-0.3 micron grains reached the coma in about 1.5 hours (1 arc sec = 740 km), were more diverse in mineralogy than the larger grains and contained crystals, and appeared to travel through the coma together. No smaller grains appeared at larger coma distances later (with slower velocities), implying that if grain fragmentation occurred, it happened within the gas acceleration zone. These results of the high spatial resolution spectroscopy (GEMINI+Michelle: Harker et 4. 2005, 2006; Subaru+COMICS: Sugita et al. 2005) revealed that the grains released from the interior were different from the nominally active areas of this comet by their: (a) crystalline content, (b) smaller size, (c) more diverse mineralogy. The temporal changes in the spectra, recorded by GEMIM+Michelle every 7 minutes, indicated that the dust mineralogy is inhomogeneous and, unexpectedly, the portion of the size distribution dominated by smaller grains has

  19. Anisotropy in the deep Earth

    Science.gov (United States)

    Romanowicz, Barbara; Wenk, Hans-Rudolf

    2017-08-01

    Seismic anisotropy has been found in many regions of the Earth's interior. Its presence in the Earth's crust has been known since the 19th century, and is due in part to the alignment of anisotropic crystals in rocks, and in part to patterns in the distribution of fractures and pores. In the upper mantle, seismic anisotropy was discovered 50 years ago, and can be attributed for the most part, to the alignment of intrinsically anisotropic olivine crystals during large scale deformation associated with convection. There is some indication for anisotropy in the transition zone, particularly in the vicinity of subducted slabs. Here we focus on the deep Earth - the lower mantle and core, where anisotropy is not yet mapped in detail, nor is there consensus on its origin. Most of the lower mantle appears largely isotropic, except in the last 200-300 km, in the D″ region, where evidence for seismic anisotropy has been accumulating since the late 1980s, mostly from shear wave splitting measurements. Recently, a picture has been emerging, where strong anisotropy is associated with high shear velocities at the edges of the large low shear velocity provinces (LLSVPs) in the central Pacific and under Africa. These observations are consistent with being due to the presence of highly anisotropic MgSiO3 post-perovskite crystals, aligned during the deformation of slabs impinging on the core-mantle boundary, and upwelling flow within the LLSVPs. We also discuss mineral physics aspects such as ultrahigh pressure deformation experiments, first principles calculations to obtain information about elastic properties, and derivation of dislocation activity based on bonding characteristics. Polycrystal plasticity simulations can predict anisotropy but models are still highly idealized and neglect the complex microstructure of polyphase aggregates with strong and weak components. A promising direction for future progress in understanding the origin of seismic anisotropy in the deep mantle

  20. [Terahertz Spectroscopic Identification with Deep Belief Network].

    Science.gov (United States)

    Ma, Shuai; Shen, Tao; Wang, Rui-qi; Lai, Hua; Yu, Zheng-tao

    2015-12-01

    Feature extraction and classification are the key issues of terahertz spectroscopy identification. Because many materials have no apparent absorption peaks in the terahertz band, it is difficult to extract theirs terahertz spectroscopy feature and identify. To this end, a novel of identify terahertz spectroscopy approach with Deep Belief Network (DBN) was studied in this paper, which combines the advantages of DBN and K-Nearest Neighbors (KNN) classifier. Firstly, cubic spline interpolation and S-G filter were used to normalize the eight kinds of substances (ATP, Acetylcholine Bromide, Bifenthrin, Buprofezin, Carbazole, Bleomycin, Buckminster and Cylotriphosphazene) terahertz transmission spectra in the range of 0.9-6 THz. Secondly, the DBN model was built by two restricted Boltzmann machine (RBM) and then trained layer by layer using unsupervised approach. Instead of using handmade features, the DBN was employed to learn suitable features automatically with raw input data. Finally, a KNN classifier was applied to identify the terahertz spectrum. Experimental results show that using the feature learned by DBN can identify the terahertz spectrum of different substances with the recognition rate of over 90%, which demonstrates that the proposed method can automatically extract the effective features of terahertz spectrum. Furthermore, this KNN classifier was compared with others (BP neural network, SOM neural network and RBF neural network). Comparisons showed that the recognition rate of KNN classifier is better than the other three classifiers. Using the approach that automatic extract terahertz spectrum features by DBN can greatly reduce the workload of feature extraction. This proposed method shows a promising future in the application of identifying the mass terahertz spectroscopy.

  1. Installation of borehole seismometer for earthquake characteristics in deep geological environment

    Energy Technology Data Exchange (ETDEWEB)

    Park, Dong Hee; Choi, Weon Hack; Cho, Sung Il; Chang, Chun Joong [KHNP CRI, Seoul (Korea, Republic of)

    2014-10-15

    Deep geological disposal is currently accepted as the most appropriate method for permanently removing spent nuclear fuel from the living sphere of humans. For implementation of deep geological disposal, we need to understand the geological changes that have taken place over the past 100,000 years, encompassing active faults, volcanic activity, elevation, ubsidence, which as yet have not been considered in assessing the site characteristics for general facilities, as well as to investigate and analyze the geological structures, fracture systems and seismic responses regarding deep geological environment about 500 meters or more underground. In regions with high seismic activity, such as Japan, the Western United States and Taiwan, borehole seismometers installed deep underground are used to monitor seismic activity during the course of seismic wave propagation at various depths and to study the stress changes due to earthquakes and analyze the connection to fault movements. Korea Hydro and Nuclear Power Co., Ltd. (KHNP) have installed the deep borehole earthquake observatory at depths of about 300 to 600 meters in order to study the seismic response characteristics in deep geological environment on June, 2014 in Andong area. This paper will show the status of deep borehole earthquake observatory and the results of background noise response characteristics of these deep borehole seismic data as a basic data analysis. We present here the status of deep borehole seismometer installation by KHNP. In order to basic data analysis for the borehole seismic observation data, this study shows the results of the orientation of seismometer and background noise characteristics by using a probability density function. Together with the ground motion data recorded by the borehole seismometers can be utilized as basic data for seismic response characteristics studies with regard to spent nuclear fuel disposal depth and as the input data for seismic hazard assessment that

  2. Discovery deep space optical communications (DSOC) transceiver

    Science.gov (United States)

    Roberts, W. Thomas

    2017-02-01

    NASA's 22 cm diameter Deep Space Optical Communications (DSOC) Transceiver is designed to provide a bidirectional optical link between a spacecraft in the inner solar system and an Earth-based optical ground station. This design, optimized for operation across a wide range of illumination conditions, is focused on minimizing blinding from stray light, and providing reliable, accurate attitude information to point its narrow communication beam accurately to the future location of the ground terminal. Though our transceiver will transmit in the 1550 nm waveband and receive in the 1064 nm waveband, the system design relies heavily on reflective optical elements, extending flexibility to be modified for use at different wavebands. The design makes use of common path propagation among transmit, receive and pointing verification optical channels to maintain precise alignment among its components, and to naturally correct for element misalignment resulting from launch or thermal element perturbations. This paper presents the results of trade studies showing the evolution of the design, unique operational characteristics of the design, elements that help to maintain minimal stray light contamination, and preliminary results from development and initial testing of a functional aluminum test model.

  3. An Unsupervised Deep Hyperspectral Anomaly Detector

    Directory of Open Access Journals (Sweden)

    Ning Ma

    2018-02-01

    Full Text Available Hyperspectral image (HSI based detection has attracted considerable attention recently in agriculture, environmental protection and military applications as different wavelengths of light can be advantageously used to discriminate different types of objects. Unfortunately, estimating the background distribution and the detection of interesting local objects is not straightforward, and anomaly detectors may give false alarms. In this paper, a Deep Belief Network (DBN based anomaly detector is proposed. The high-level features and reconstruction errors are learned through the network in a manner which is not affected by previous background distribution assumption. To reduce contamination by local anomalies, adaptive weights are constructed from reconstruction errors and statistical information. By using the code image which is generated during the inference of DBN and modified by adaptively updated weights, a local Euclidean distance between under test pixels and their neighboring pixels is used to determine the anomaly targets. Experimental results on synthetic and recorded HSI datasets show the performance of proposed method outperforms the classic global Reed-Xiaoli detector (RXD, local RX detector (LRXD and the-state-of-the-art Collaborative Representation detector (CRD.

  4. Folding Membrane Proteins by Deep Transfer Learning

    KAUST Repository

    Wang, Sheng

    2017-08-29

    Computational elucidation of membrane protein (MP) structures is challenging partially due to lack of sufficient solved structures for homology modeling. Here, we describe a high-throughput deep transfer learning method that first predicts MP contacts by learning from non-MPs and then predicts 3D structure models using the predicted contacts as distance restraints. Tested on 510 non-redundant MPs, our method has contact prediction accuracy at least 0.18 better than existing methods, predicts correct folds for 218 MPs, and generates 3D models with root-mean-square deviation (RMSD) less than 4 and 5 Å for 57 and 108 MPs, respectively. A rigorous blind test in the continuous automated model evaluation project shows that our method predicted high-resolution 3D models for two recent test MPs of 210 residues with RMSD ∼2 Å. We estimated that our method could predict correct folds for 1,345–1,871 reviewed human multi-pass MPs including a few hundred new folds, which shall facilitate the discovery of drugs targeting at MPs.

  5. Localization noise in deep subwavelength plasmonic devices

    Science.gov (United States)

    Ghoreyshi, Ali; Victora, R. H.

    2018-05-01

    The grain shape dependence of absorption has been investigated in metal-insulator thin films. We demonstrate that randomness in the size and shape of plasmonic particles can lead to Anderson localization of polarization modes in the deep subwavelength regime. These localized modes can contribute to significant variation in the local field. In the case of plasmonic nanodevices, the effects of the localized modes have been investigated by mapping an electrostatic Hamiltonian onto the Anderson Hamiltonian in the presence of a random vector potential. We show that local behavior of the optical beam can be understood in terms of the weighted local density of the localized modes of the depolarization field. Optical nanodevices that operate on a length scale with high variation in the density of states of localized modes will experience a previously unidentified localized noise. This localization noise contributes uncertainty to the output of plasmonic nanodevices and limits their scalability. In particular, the resulting impact on heat-assisted magnetic recording is discussed.

  6. Quantum effects in deep inelastic neutron scattering

    International Nuclear Information System (INIS)

    Mayers, J.

    1989-07-01

    In the Impulse Approximation (IA), which is used to interpret deep inelastic neutron scattering (DINS) measurements, it is assumed both that the target system can be treated as a gas of free atoms and that the struck atom recoils freely after the collision with the neutron. Departures from the IA are generally attributed to final state effects (FSE), which are due to the inaccuracy of the latter assumption. However it is shown that even when FSE are neglected, significant departures from the IA occur at low temperatures due to inaccuracies in the former assumption. These are referred to as initial state effects (ISE) and are due to the quantum nature of the initial state. Comparison with experimental data and exactly soluble models shows that ISE largely account for observed asymmetries and peak shifts in the neutron scattering function S(q,ω), compared with the IA prediction. It is shown that when FSE are neglected, ISE can also be neglected when either the momentum transfer or the temperature is high. Finally it is shown that FSE should be negligible at high momentum transfers in systems other than quantum fluids and that therefore in this regime the IA is reached in such systems. (author)

  7. Habitat Concepts for Deep Space Exploration

    Science.gov (United States)

    Smitherman, David; Griffin, Brand N.

    2014-01-01

    Future missions under consideration requiring human habitation beyond the International Space Station (ISS) include deep space habitats in the lunar vicinity to support asteroid retrieval missions, human and robotic lunar missions, satellite servicing, and Mars vehicle servicing missions. Habitat designs are also under consideration for missions beyond the Earth-Moon system, including transfers to near-Earth asteroids and Mars orbital destinations. A variety of habitat layouts have been considered, including those derived from the existing ISS designs and those that could be fabricated from the Space Launch System (SLS) propellant tanks. This paper presents a comparison showing several options for asteroid, lunar, and Mars mission habitats using ISS derived and SLS derived modules and identifies some of the advantages and disadvantages inherent in each. Key findings indicate that the larger SLS diameter modules offer built-in compatibility with the launch vehicle, single launch capability without on-orbit assembly, improved radiation protection, lighter structures per unit volume, and sufficient volume to accommodate consumables for long duration missions without resupply. The information provided with the findings includes mass and volume comparison data that should be helpful to future exploration mission planning efforts.

  8. Autofocusing in digital holography using deep learning

    Science.gov (United States)

    Ren, Zhenbo; Xu, Zhimin; Lam, Edmund Y.

    2018-02-01

    In digital holography, it is critical to know the distance in order to reconstruct the multi-sectional object. This autofocusing is traditionally solved by reconstructing a stack of in-focus and out-of-focus images and using some focus metric, such as entropy or variance, to calculate the sharpness of each reconstructed image. Then the distance corresponding to the sharpest image is determined as the focal position. This method is effective but computationally demanding and time-consuming. To get an accurate estimation, one has to reconstruct many images. Sometimes after a coarse search, a refinement is needed. To overcome this problem in autofocusing, we propose to use deep learning, i.e., a convolutional neural network (CNN), to solve this problem. Autofocusing is viewed as a classification problem, in which the true distance is transferred as a label. To estimate the distance is equated to labeling a hologram correctly. To train such an algorithm, totally 1000 holograms are captured under the same environment, i.e., exposure time, incident angle, object, except the distance. There are 5 labels corresponding to 5 distances. These data are randomly split into three datasets to train, validate and test a CNN network. Experimental results show that the trained network is capable of predicting the distance without reconstructing or knowing any physical parameters about the setup. The prediction time using this method is far less than traditional autofocusing methods.

  9. Nuclear wastes beneath the deep sea floor

    International Nuclear Information System (INIS)

    Bishop, W.P.; Hollister, C.D.

    1974-01-01

    Projections of energy demands for the year 2000 show that nuclear power will likely be one of our energy sources. But the benefits of nuclear power must be balanced against the drawbacks of its by-product: high-level wastes. While it may become possible to completely destroy or eliminate these wastes, it is at least equally possible that we may have to dispose of them on earth in such a way as to assure their isolation from man for periods of the order of a million years. Undersea regions in the middle of tectonic plates and in the approximate center of major current gyres offer some conceptual promise for waste disposal because of their geologic stability and comparatively low organic productivity. The advantages of this concept and the types of detailed information needed for its accurate assessment are discussed. The technical feasibility of permanent disposal beneath the deep sea floor cannot be accurately assessed with present knowledge, and there is a need for a thorough study of the types and rates of processes that affect this part of the earth's surface. Basic oceanographic research aimed at understanding these processes is yielding answers that apply to this societal need. (U.S.)

  10. Obstetrical complications of endometriosis, particularly deep endometriosis.

    Science.gov (United States)

    Leone Roberti Maggiore, Umberto; Inversetti, Annalisa; Schimberni, Matteo; Viganò, Paola; Giorgione, Veronica; Candiani, Massimo

    2017-12-01

    Over the past few years, a new topic in the field of endometriosis has emerged: the potential impact of the disease on pregnancy outcomes. This review aims to summarize in detail the available evidence on the relationship between endometriosis, particularly deep endometriosis (DE), and obstetrical outcomes. Acute complications of DE, such as spontaneous hemoperitoneum, bowel perforation, and uterine rupture, may occur during pregnancy. Although these events represent life-threatening conditions, they are rare and unpredictable. Therefore, the current literature does not support any kind of prophylactic surgery before pregnancy to prevent such complications. Results on the impact of DE on obstetrical outcomes are debatable and characterized by several limitations, including small sample size, lack of adjustment for confounders, lack of adequate control subjects, and other methodologic flaws. For these reasons, it is not possible to draw conclusions on this topic. The strongest evidence shows that DE is associated with higher rates of placenta previa; for other obstetrical outcomes, such as miscarriage, intrauterine growth restriction, preterm birth and hypertensive disorders, results are controversial. Although it is unlikely that surgery of DE may modify the impact of the disease on the course of pregnancy, no study has yet investigated this issue. Copyright © 2017 American Society for Reproductive Medicine. Published by Elsevier Inc. All rights reserved.

  11. Deep image mining for diabetic retinopathy screening.

    Science.gov (United States)

    Quellec, Gwenolé; Charrière, Katia; Boudi, Yassine; Cochener, Béatrice; Lamard, Mathieu

    2017-07-01

    Deep learning is quickly becoming the leading methodology for medical image analysis. Given a large medical archive, where each image is associated with a diagnosis, efficient pathology detectors or classifiers can be trained with virtually no expert knowledge about the target pathologies. However, deep learning algorithms, including the popular ConvNets, are black boxes: little is known about the local patterns analyzed by ConvNets to make a decision at the image level. A solution is proposed in this paper to create heatmaps showing which pixels in images play a role in the image-level predictions. In other words, a ConvNet trained for image-level classification can be used to detect lesions as well. A generalization of the backpropagation method is proposed in order to train ConvNets that produce high-quality heatmaps. The proposed solution is applied to diabetic retinopathy (DR) screening in a dataset of almost 90,000 fundus photographs from the 2015 Kaggle Diabetic Retinopathy competition and a private dataset of almost 110,000 photographs (e-ophtha). For the task of detecting referable DR, very good detection performance was achieved: A z =0.954 in Kaggle's dataset and A z =0.949 in e-ophtha. Performance was also evaluated at the image level and at the lesion level in the DiaretDB1 dataset, where four types of lesions are manually segmented: microaneurysms, hemorrhages, exudates and cotton-wool spots. For the task of detecting images containing these four lesion types, the proposed detector, which was trained to detect referable DR, outperforms recent algorithms trained to detect those lesions specifically, with pixel-level supervision. At the lesion level, the proposed detector outperforms heatmap generation algorithms for ConvNets. This detector is part of the Messidor® system for mobile eye pathology screening. Because it does not rely on expert knowledge or manual segmentation for detecting relevant patterns, the proposed solution is a promising image

  12. Sunspot drawings handwritten character recognition method based on deep learning

    Science.gov (United States)

    Zheng, Sheng; Zeng, Xiangyun; Lin, Ganghua; Zhao, Cui; Feng, Yongli; Tao, Jinping; Zhu, Daoyuan; Xiong, Li

    2016-05-01

    High accuracy scanned sunspot drawings handwritten characters recognition is an issue of critical importance to analyze sunspots movement and store them in the database. This paper presents a robust deep learning method for scanned sunspot drawings handwritten characters recognition. The convolution neural network (CNN) is one algorithm of deep learning which is truly successful in training of multi-layer network structure. CNN is used to train recognition model of handwritten character images which are extracted from the original sunspot drawings. We demonstrate the advantages of the proposed method on sunspot drawings provided by Chinese Academy Yunnan Observatory and obtain the daily full-disc sunspot numbers and sunspot areas from the sunspot drawings. The experimental results show that the proposed method achieves a high recognition accurate rate.

  13. Realization of Chinese word segmentation based on deep learning method

    Science.gov (United States)

    Wang, Xuefei; Wang, Mingjiang; Zhang, Qiquan

    2017-08-01

    In recent years, with the rapid development of deep learning, it has been widely used in the field of natural language processing. In this paper, I use the method of deep learning to achieve Chinese word segmentation, with large-scale corpus, eliminating the need to construct additional manual characteristics. In the process of Chinese word segmentation, the first step is to deal with the corpus, use word2vec to get word embedding of the corpus, each character is 50. After the word is embedded, the word embedding feature is fed to the bidirectional LSTM, add a linear layer to the hidden layer of the output, and then add a CRF to get the model implemented in this paper. Experimental results show that the method used in the 2014 People's Daily corpus to achieve a satisfactory accuracy.

  14. Transcriptome sequences resolve deep relationships of the grape family.

    Science.gov (United States)

    Wen, Jun; Xiong, Zhiqiang; Nie, Ze-Long; Mao, Likai; Zhu, Yabing; Kan, Xian-Zhao; Ickert-Bond, Stefanie M; Gerrath, Jean; Zimmer, Elizabeth A; Fang, Xiao-Dong

    2013-01-01

    Previous phylogenetic studies of the grape family (Vitaceae) yielded poorly resolved deep relationships, thus impeding our understanding of the evolution of the family. Next-generation sequencing now offers access to protein coding sequences very easily, quickly and cost-effectively. To improve upon earlier work, we extracted 417 orthologous single-copy nuclear genes from the transcriptomes of 15 species of the Vitaceae, covering its phylogenetic diversity. The resulting transcriptome phylogeny provides robust support for the deep relationships, showing the phylogenetic utility of transcriptome data for plants over a time scale at least since the mid-Cretaceous. The pros and cons of transcriptome data for phylogenetic inference in plants are also evaluated.

  15. Random Deep Belief Networks for Recognizing Emotions from Speech Signals.

    Science.gov (United States)

    Wen, Guihua; Li, Huihui; Huang, Jubing; Li, Danyang; Xun, Eryang

    2017-01-01

    Now the human emotions can be recognized from speech signals using machine learning methods; however, they are challenged by the lower recognition accuracies in real applications due to lack of the rich representation ability. Deep belief networks (DBN) can automatically discover the multiple levels of representations in speech signals. To make full of its advantages, this paper presents an ensemble of random deep belief networks (RDBN) method for speech emotion recognition. It firstly extracts the low level features of the input speech signal and then applies them to construct lots of random subspaces. Each random subspace is then provided for DBN to yield the higher level features as the input of the classifier to output an emotion label. All outputted emotion labels are then fused through the majority voting to decide the final emotion label for the input speech signal. The conducted experimental results on benchmark speech emotion databases show that RDBN has better accuracy than the compared methods for speech emotion recognition.

  16. Transcriptome sequences resolve deep relationships of the grape family.

    Directory of Open Access Journals (Sweden)

    Jun Wen

    Full Text Available Previous phylogenetic studies of the grape family (Vitaceae yielded poorly resolved deep relationships, thus impeding our understanding of the evolution of the family. Next-generation sequencing now offers access to protein coding sequences very easily, quickly and cost-effectively. To improve upon earlier work, we extracted 417 orthologous single-copy nuclear genes from the transcriptomes of 15 species of the Vitaceae, covering its phylogenetic diversity. The resulting transcriptome phylogeny provides robust support for the deep relationships, showing the phylogenetic utility of transcriptome data for plants over a time scale at least since the mid-Cretaceous. The pros and cons of transcriptome data for phylogenetic inference in plants are also evaluated.

  17. Head pose estimation algorithm based on deep learning

    Science.gov (United States)

    Cao, Yuanming; Liu, Yijun

    2017-05-01

    Head pose estimation has been widely used in the field of artificial intelligence, pattern recognition and intelligent human-computer interaction and so on. Good head pose estimation algorithm should deal with light, noise, identity, shelter and other factors robustly, but so far how to improve the accuracy and robustness of attitude estimation remains a major challenge in the field of computer vision. A method based on deep learning for pose estimation is presented. Deep learning with a strong learning ability, it can extract high-level image features of the input image by through a series of non-linear operation, then classifying the input image using the extracted feature. Such characteristics have greater differences in pose, while they are robust of light, identity, occlusion and other factors. The proposed head pose estimation is evaluated on the CAS-PEAL data set. Experimental results show that this method is effective to improve the accuracy of pose estimation.

  18. Using Supervised Deep Learning for Human Age Estimation Problem

    Science.gov (United States)

    Drobnyh, K. A.; Polovinkin, A. N.

    2017-05-01

    Automatic facial age estimation is a challenging task upcoming in recent years. In this paper, we propose using the supervised deep learning features to improve an accuracy of the existing age estimation algorithms. There are many approaches solving the problem, an active appearance model and the bio-inspired features are two of them which showed the best accuracy. For experiments we chose popular publicly available FG-NET database, which contains 1002 images with a broad variety of light, pose, and expression. LOPO (leave-one-person-out) method was used to estimate the accuracy. Experiments demonstrated that adding supervised deep learning features has improved accuracy for some basic models. For example, adding the features to an active appearance model gave the 4% gain (the error decreased from 4.59 to 4.41).

  19. Clustered deep shadow maps for integrated polyhedral and volume rendering

    KAUST Repository

    Bornik, Alexander

    2012-01-01

    This paper presents a hardware-accelerated approach for shadow computation in scenes containing both complex volumetric objects and polyhedral models. Our system is the first hardware accelerated complete implementation of deep shadow maps, which unifies the computation of volumetric and geometric shadows. Up to now such unified computation was limited to software-only rendering . Previous hardware accelerated techniques can handle only geometric or only volumetric scenes - both resulting in the loss of important properties of the original concept. Our approach supports interactive rendering of polyhedrally bounded volumetric objects on the GPU based on ray casting. The ray casting can be conveniently used for both the shadow map computation and the rendering. We show how anti-aliased high-quality shadows are feasible in scenes composed of multiple overlapping translucent objects, and how sparse scenes can be handled efficiently using clustered deep shadow maps. © 2012 Springer-Verlag.

  20. Cyanide Suicide After Deep Web Shopping: A Case Report.

    Science.gov (United States)

    Le Garff, Erwan; Delannoy, Yann; Mesli, Vadim; Allorge, Delphine; Hédouin, Valéry; Tournel, Gilles

    2016-09-01

    Cyanide is a product that is known for its use in industrial or laboratory processes, as well as for intentional intoxication. The toxicity of cyanide is well described in humans with rapid inhibition of cellular aerobic metabolism after ingestion or inhalation, leading to severe clinical effects that are frequently lethal. We report the case of a young white man found dead in a hotel room after self-poisoning with cyanide ordered in the deep Web. This case shows a probable complex suicide kit use including cyanide, as a lethal tool, and dextromethorphan, as a sedative and anxiolytic substance. This case is an original example of the emerging deep Web shopping in illegal drug procurement.

  1. Deep-Learning-Based Approach for Prediction of Algal Blooms

    Directory of Open Access Journals (Sweden)

    Feng Zhang

    2016-10-01

    Full Text Available Algal blooms have recently become a critical global environmental concern which might put economic development and sustainability at risk. However, the accurate prediction of algal blooms remains a challenging scientific problem. In this study, a novel prediction approach for algal blooms based on deep learning is presented—a powerful tool to represent and predict highly dynamic and complex phenomena. The proposed approach constructs a five-layered model to extract detailed relationships between the density of phytoplankton cells and various environmental parameters. The algal blooms can be predicted by the phytoplankton density obtained from the output layer. A case study is conducted in coastal waters of East China using both our model and a traditional back-propagation neural network for comparison. The results show that the deep-learning-based model yields better generalization and greater accuracy in predicting algal blooms than a traditional shallow neural network does.

  2. DeepDive: Declarative Knowledge Base Construction.

    Science.gov (United States)

    De Sa, Christopher; Ratner, Alex; Ré, Christopher; Shin, Jaeho; Wang, Feiran; Wu, Sen; Zhang, Ce

    2016-03-01

    The dark data extraction or knowledge base construction (KBC) problem is to populate a SQL database with information from unstructured data sources including emails, webpages, and pdf reports. KBC is a long-standing problem in industry and research that encompasses problems of data extraction, cleaning, and integration. We describe DeepDive, a system that combines database and machine learning ideas to help develop KBC systems. The key idea in DeepDive is that statistical inference and machine learning are key tools to attack classical data problems in extraction, cleaning, and integration in a unified and more effective manner. DeepDive programs are declarative in that one cannot write probabilistic inference algorithms; instead, one interacts by defining features or rules about the domain. A key reason for this design choice is to enable domain experts to build their own KBC systems. We present the applications, abstractions, and techniques of DeepDive employed to accelerate construction of KBC systems.

  3. Variational inference & deep learning : A new synthesis

    NARCIS (Netherlands)

    Kingma, D.P.

    2017-01-01

    In this thesis, Variational Inference and Deep Learning: A New Synthesis, we propose novel solutions to the problems of variational (Bayesian) inference, generative modeling, representation learning, semi-supervised learning, and stochastic optimization.

  4. Variational inference & deep learning: A new synthesis

    OpenAIRE

    Kingma, D.P.

    2017-01-01

    In this thesis, Variational Inference and Deep Learning: A New Synthesis, we propose novel solutions to the problems of variational (Bayesian) inference, generative modeling, representation learning, semi-supervised learning, and stochastic optimization.

  5. DNA Replication Profiling Using Deep Sequencing.

    Science.gov (United States)

    Saayman, Xanita; Ramos-Pérez, Cristina; Brown, Grant W

    2018-01-01

    Profiling of DNA replication during progression through S phase allows a quantitative snap-shot of replication origin usage and DNA replication fork progression. We present a method for using deep sequencing data to profile DNA replication in S. cerevisiae.

  6. Evaluation of static resistance of deep foundations.

    Science.gov (United States)

    2017-05-01

    The focus of this research was to evaluate and improve Florida Department of Transportation (FDOT) FB-Deep software prediction of nominal resistance of H-piles, prestressed concrete piles in limestone, large diameter (> 36) open steel and concrete...

  7. The deep ocean under climate change.

    Science.gov (United States)

    Levin, Lisa A; Le Bris, Nadine

    2015-11-13

    The deep ocean absorbs vast amounts of heat and carbon dioxide, providing a critical buffer to climate change but exposing vulnerable ecosystems to combined stresses of warming, ocean acidification, deoxygenation, and altered food inputs. Resulting changes may threaten biodiversity and compromise key ocean services that maintain a healthy planet and human livelihoods. There exist large gaps in understanding of the physical and ecological feedbacks that will occur. Explicit recognition of deep-ocean climate mitigation and inclusion in adaptation planning by the United Nations Framework Convention on Climate Change (UNFCCC) could help to expand deep-ocean research and observation and to protect the integrity and functions of deep-ocean ecosystems. Copyright © 2015, American Association for the Advancement of Science.

  8. Deep gold mine fracture zone behaviour

    CSIR Research Space (South Africa)

    Napier, JAL

    1998-12-01

    Full Text Available The investigation of the behaviour of the fracture zone surrounding deep level gold mine stopes is detailed in three main sections of this report. Section 2 outlines the ongoing study of fundamental fracture process and their numerical...

  9. Deep Ultraviolet Macroporous Silicon Filters, Phase I

    Data.gov (United States)

    National Aeronautics and Space Administration — This SBIR Phase I proposal describes a novel method to make deep and far UV optical filters from macroporous silicon. This type of filter consists of an array of...

  10. Toolkits and Libraries for Deep Learning.

    Science.gov (United States)

    Erickson, Bradley J; Korfiatis, Panagiotis; Akkus, Zeynettin; Kline, Timothy; Philbrick, Kenneth

    2017-08-01

    Deep learning is an important new area of machine learning which encompasses a wide range of neural network architectures designed to complete various tasks. In the medical imaging domain, example tasks include organ segmentation, lesion detection, and tumor classification. The most popular network architecture for deep learning for images is the convolutional neural network (CNN). Whereas traditional machine learning requires determination and calculation of features from which the algorithm learns, deep learning approaches learn the important features as well as the proper weighting of those features to make predictions for new data. In this paper, we will describe some of the libraries and tools that are available to aid in the construction and efficient execution of deep learning as applied to medical images.

  11. Deep-Sea Soft Coral Habitat Suitability

    Data.gov (United States)

    National Oceanic and Atmospheric Administration, Department of Commerce — Deep-sea corals, also known as cold water corals, create complex communities that provide habitat for a variety of invertebrate and fish species, such as grouper,...

  12. Photon diffractive dissociation in deep inelastic scattering

    International Nuclear Information System (INIS)

    Ryskin, M.G.

    1990-01-01

    The new ep-collider HERA gives us the possibility to study the diffractive dissociation of virtual photon in deep inelastic ep-collision. The process of photon dissociation in deep inelastic scattering is the most direct way to measure the value of triple-pomeron vertex G 3P . It was shown that the value of the correct bare vertex G 3P may more than 4 times exceeds its effective value measuring in the triple-reggeon region and reaches the value of about 40-50% of the elastic pp-pomeron vertex. On the contrary in deep inelastic processes the perpendicular momenta q t of the secondary particles are large enough. Thus in deep inelastic reactions one can measure the absolute value of G 3P vertex in the most direct way and compare its value and q t dependence with the leading log QCD predictions

  13. Applications of Deep Learning in Biomedicine.

    Science.gov (United States)

    Mamoshina, Polina; Vieira, Armando; Putin, Evgeny; Zhavoronkov, Alex

    2016-05-02

    Increases in throughput and installed base of biomedical research equipment led to a massive accumulation of -omics data known to be highly variable, high-dimensional, and sourced from multiple often incompatible data platforms. While this data may be useful for biomarker identification and drug discovery, the bulk of it remains underutilized. Deep neural networks (DNNs) are efficient algorithms based on the use of compositional layers of neurons, with advantages well matched to the challenges -omics data presents. While achieving state-of-the-art results and even surpassing human accuracy in many challenging tasks, the adoption of deep learning in biomedicine has been comparatively slow. Here, we discuss key features of deep learning that may give this approach an edge over other machine learning methods. We then consider limitations and review a number of applications of deep learning in biomedical studies demonstrating proof of concept and practical utility.

  14. Mean associative multiplicities in deep inelastic processes

    International Nuclear Information System (INIS)

    Dzhaparidze, G.Sh.; Kiselev, A.V.; Petrov, V.A.

    1981-01-01

    The associative hadron multiplicities in deep inelastic and Drell--Yan processes are studied. In particular the mean multiplicities in different hard processes in QCD are found to be determined by the mean multiplicity in parton jet [ru

  15. Deep-Sea Stony Coral Habitat Suitability

    Data.gov (United States)

    National Oceanic and Atmospheric Administration, Department of Commerce — Deep-sea corals, also known as cold water corals, create complex communities that provide habitat for a variety of invertebrate and fish species, such as grouper,...

  16. Deep Learning and Applications in Computational Biology

    KAUST Repository

    Zeng, Jianyang

    2016-01-01

    In this work, we develop a general and flexible deep learning framework for modeling structural binding preferences and predicting binding sites of RBPs, which takes (predicted) RNA tertiary structural information

  17. Leading particle in deep inelastic scattering

    International Nuclear Information System (INIS)

    Petrov, V.A.

    1984-01-01

    The leading particle effect in deep inelastic scattering is considered. The change of the characteris cs shape of the leading particle inclusive spectrum with Q 2 is estimated to be rather significant at very high Q 2

  18. Progress in deep-UV photoresists

    Indian Academy of Sciences (India)

    Unknown

    This paper reviews the recent development and challenges of deep-UV photoresists and their ... small amount of acid, when exposed to light by photo- chemical ... anomalous insoluble skin and linewidth shift when the. PEB was delayed.

  19. Methods in mooring deep sea sediment traps

    Digital Repository Service at National Institute of Oceanography (India)

    Venkatesan, R.; Fernando, V.; Rajaraman, V.S.; Janakiraman, G.

    The experience gained during the process of deployment and retrieval of nearly 39 sets of deep sea sediment trap moorings on various ships like FS Sonne, ORV Sagarkanya and DSV Nand Rachit are outlined. The various problems encountered...

  20. Deep Water Horizon (HB1006, EK60)

    Data.gov (United States)

    National Oceanic and Atmospheric Administration, Department of Commerce — Monitor and measure the biological, chemical, and physical environment in the area of the oil spill from the deep water horizon oil rig in the Gulf of Mexico. A wide...

  1. Genetic diversity of archaea in deep-sea hydrothermal vent environments.

    OpenAIRE

    Takai, K; Horikoshi, K

    1999-01-01

    Molecular phylogenetic analysis of naturally occurring archaeal communities in deep-sea hydrothermal vent environments was carried out by PCR-mediated small subunit rRNA gene (SSU rDNA) sequencing. As determined through partial sequencing of rDNA clones amplified with archaea-specific primers, the archaeal populations in deep-sea hydrothermal vent environments showed a great genetic diversity, and most members of these populations appeared to be uncultivated and unidentified organisms. In the...

  2. DeepRain: ConvLSTM Network for Precipitation Prediction using Multichannel Radar Data

    OpenAIRE

    Kim, Seongchan; Hong, Seungkyun; Joh, Minsu; Song, Sa-kwang

    2017-01-01

    Accurate rainfall forecasting is critical because it has a great impact on people's social and economic activities. Recent trends on various literatures show that Deep Learning (Neural Network) is a promising methodology to tackle many challenging tasks. In this study, we introduce a brand-new data-driven precipitation prediction model called DeepRain. This model predicts the amount of rainfall from weather radar data, which is three-dimensional and four-channel data, using convolutional LSTM...

  3. Discrimination and competition between complete fusion and deep inelastic reactions induced by heavy ions

    International Nuclear Information System (INIS)

    Hanappe, F.; Tamain, B.

    1977-01-01

    One tries to find a way to discriminate between fission following fusion and deep inelastic processes with large mass transfer. Fragment analysis (kinetic energy, mass, charge distributions) gives generally no answer. The deexcitation properties of the fragments (gamma ray, charged particles and neutron emission) are difficult to interpret, and only recent results concerning neutron emission show different patterns for both processes. The reasons for which a system evolves towards deep inelastic processes rather than fusion are discussed

  4. Pengembangan Pendidikan Karakter Dalam Mata Kuliah Evaluasi Pembelajaran Melalui Pendekatan Deep Approach to Learning

    OpenAIRE

    Suryani, Nanik; Pramushinto, Hengky

    2012-01-01

    The objectives of this study are to find and to test the model of characters education in Learning Evaluation Subject through deep approach to learning. The subject of the study is the class of Learning Evaluation of Office Administration Program, Economics Education Department, Economics Faculty, Semarang State University. The data are collected by a test, and then analyzed by qualitative descriptive. The result of this study showed that the model of characters education through deep approac...

  5. Time-lagged autoencoders: Deep learning of slow collective variables for molecular kinetics

    Science.gov (United States)

    Wehmeyer, Christoph; Noé, Frank

    2018-06-01

    Inspired by the success of deep learning techniques in the physical and chemical sciences, we apply a modification of an autoencoder type deep neural network to the task of dimension reduction of molecular dynamics data. We can show that our time-lagged autoencoder reliably finds low-dimensional embeddings for high-dimensional feature spaces which capture the slow dynamics of the underlying stochastic processes—beyond the capabilities of linear dimension reduction techniques.

  6. Fossil manganese nodules from Timor: geochemical and radiochemical evidence for deep-sea origin

    International Nuclear Information System (INIS)

    Margolis, S.V.; Fein, C.D.; Glasby, G.P.; Audley-Charles, M.G.

    1978-01-01

    Fossil Mn nodules of Cretaceous age from western Timor exhibit chemical, structural and radioisotope compositions consistent with their being of deep-sea origin. These nodules show characteristics similar to nodules now found at depths of 3,500-5,000 m in the Pacific and Indian Oceans. Slight differences in the fine structure and chemistry of these nodules and modern deep-sea nodules are attributed to diagenetic alteration after uplift of enclosing sediments

  7. Results of testing the E9 multiple probe lateral logging device in deep wells in the eastern Pre-caucasus

    Energy Technology Data Exchange (ETDEWEB)

    Boyarchuk, A.F.; Kochetkov, V.T.; Kucherov, R.A.

    1981-07-01

    The integrated lateral logging device E9 developed for investigating deep and extra-deep wells, permitting measurement of apparent resistances by three probes at different depths, is described. It is heat and pressure resistant (up to 200/degree/C, 120 MPa). The tests showed that under certain favorable conditions the device is fairly effective.

  8. Biodiversity loss from deep-sea mining

    OpenAIRE

    C. L. Van Dover; J. A. Ardron; E. Escobar; M. Gianni; K. M. Gjerde; A. Jaeckel; D. O. B. Jones; L. A. Levin; H. Niner; L. Pendleton; C. R. Smith; T. Thiele; P. J. Turner; L. Watling; P. P. E. Weaver

    2017-01-01

    The emerging deep-sea mining industry is seen by some to be an engine for economic development in the maritime sector. The International Seabed Authority (ISA) – the body that regulates mining activities on the seabed beyond national jurisdiction – must also protect the marine environment from harmful effects that arise from mining. The ISA is currently drafting a regulatory framework for deep-sea mining that includes measures for environmental protection. Responsible mining increasingly stri...

  9. DEEP VADOSE ZONE TREATABILITY TEST PLAN

    International Nuclear Information System (INIS)

    Chronister, G.B.; Truex, M.J.

    2009-01-01

    (sm b ullet) Treatability test plan published in 2008 (sm b ullet) Outlines technology treatability activities for evaluating application of in situ technologies and surface barriers to deep vadose zone contamination (technetium and uranium) (sm b ullet) Key elements - Desiccation testing - Testing of gas-delivered reactants for in situ treatment of uranium - Evaluating surface barrier application to deep vadose zone - Evaluating in situ grouting and soil flushing

  10. Deep inelastic inclusive weak and electromagnetic interactions

    International Nuclear Information System (INIS)

    Adler, S.L.

    1976-01-01

    The theory of deep inelastic inclusive interactions is reviewed, emphasizing applications to electromagnetic and weak charged current processes. The following reactions are considered: e + N → e + X, ν + N → μ - + X, anti ν + N → μ + + X where X denotes a summation over all final state hadrons and the ν's are muon neutrinos. After a discussion of scaling, the quark-parton model is invoked to explain the principle experimental features of deep inelastic inclusive reactions

  11. Short-term Memory of Deep RNN

    OpenAIRE

    Gallicchio, Claudio

    2018-01-01

    The extension of deep learning towards temporal data processing is gaining an increasing research interest. In this paper we investigate the properties of state dynamics developed in successive levels of deep recurrent neural networks (RNNs) in terms of short-term memory abilities. Our results reveal interesting insights that shed light on the nature of layering as a factor of RNN design. Noticeably, higher layers in a hierarchically organized RNN architecture results to be inherently biased ...

  12. Deep Learning for Video Game Playing

    OpenAIRE

    Justesen, Niels; Bontrager, Philip; Togelius, Julian; Risi, Sebastian

    2017-01-01

    In this article, we review recent Deep Learning advances in the context of how they have been applied to play different types of video games such as first-person shooters, arcade games, and real-time strategy games. We analyze the unique requirements that different game genres pose to a deep learning system and highlight important open challenges in the context of applying these machine learning methods to video games, such as general game playing, dealing with extremely large decision spaces...

  13. Life Support for Deep Space and Mars

    Science.gov (United States)

    Jones, Harry W.; Hodgson, Edward W.; Kliss, Mark H.

    2014-01-01

    How should life support for deep space be developed? The International Space Station (ISS) life support system is the operational result of many decades of research and development. Long duration deep space missions such as Mars have been expected to use matured and upgraded versions of ISS life support. Deep space life support must use the knowledge base incorporated in ISS but it must also meet much more difficult requirements. The primary new requirement is that life support in deep space must be considerably more reliable than on ISS or anywhere in the Earth-Moon system, where emergency resupply and a quick return are possible. Due to the great distance from Earth and the long duration of deep space missions, if life support systems fail, the traditional approaches for emergency supply of oxygen and water, emergency supply of parts, and crew return to Earth or escape to a safe haven are likely infeasible. The Orbital Replacement Unit (ORU) maintenance approach used by ISS is unsuitable for deep space with ORU's as large and complex as those originally provided in ISS designs because it minimizes opportunities for commonality of spares, requires replacement of many functional parts with each failure, and results in substantial launch mass and volume penalties. It has become impractical even for ISS after the shuttle era, resulting in the need for ad hoc repair activity at lower assembly levels with consequent crew time penalties and extended repair timelines. Less complex, more robust technical approaches may be needed to meet the difficult deep space requirements for reliability, maintainability, and reparability. Developing an entirely new life support system would neglect what has been achieved. The suggested approach is use the ISS life support technologies as a platform to build on and to continue to improve ISS subsystems while also developing new subsystems where needed to meet deep space requirements.

  14. Deep Predictive Models in Interactive Music

    OpenAIRE

    Martin, Charles P.; Ellefsen, Kai Olav; Torresen, Jim

    2018-01-01

    Automatic music generation is a compelling task where much recent progress has been made with deep learning models. In this paper, we ask how these models can be integrated into interactive music systems; how can they encourage or enhance the music making of human users? Musical performance requires prediction to operate instruments, and perform in groups. We argue that predictive models could help interactive systems to understand their temporal context, and ensemble behaviour. Deep learning...

  15. Predicting Process Behaviour using Deep Learning

    OpenAIRE

    Evermann, Joerg; Rehse, Jana-Rebecca; Fettke, Peter

    2016-01-01

    Predicting business process behaviour is an important aspect of business process management. Motivated by research in natural language processing, this paper describes an application of deep learning with recurrent neural networks to the problem of predicting the next event in a business process. This is both a novel method in process prediction, which has largely relied on explicit process models, and also a novel application of deep learning methods. The approach is evaluated on two real da...

  16. A Deep Learning Approach to Drone Monitoring

    OpenAIRE

    Chen, Yueru; Aggarwal, Pranav; Choi, Jongmoo; Kuo, C. -C. Jay

    2017-01-01

    A drone monitoring system that integrates deep-learning-based detection and tracking modules is proposed in this work. The biggest challenge in adopting deep learning methods for drone detection is the limited amount of training drone images. To address this issue, we develop a model-based drone augmentation technique that automatically generates drone images with a bounding box label on drone's location. To track a small flying drone, we utilize the residual information between consecutive i...

  17. Bank of Weight Filters for Deep CNNs

    Science.gov (United States)

    2016-11-22

    very large even on the best available hardware . In some studies in transfer learning it has been observed that the network learnt on one task can be...CNNs. Keywords: CNN, deep learning , neural networks, transfer learning , bank of weigh filters, BWF 1. Introduction Object recognition is an important...of CNNs (or, in general, of deep neural networks) is that feature generation part is fused with the classifier part and both parts are learned together

  18. Leveraging multiple datasets for deep leaf counting

    OpenAIRE

    Dobrescu, Andrei; Giuffrida, Mario Valerio; Tsaftaris, Sotirios A

    2017-01-01

    The number of leaves a plant has is one of the key traits (phenotypes) describing its development and growth. Here, we propose an automated, deep learning based approach for counting leaves in model rosette plants. While state-of-the-art results on leaf counting with deep learning methods have recently been reported, they obtain the count as a result of leaf segmentation and thus require per-leaf (instance) segmentation to train the models (a rather strong annotation). Instead, our method tre...

  19. Analysis for Behavior of Reinforcement Lap Splices in Deep Beams

    Directory of Open Access Journals (Sweden)

    Ammar Yaser Ali

    2018-03-01

    Full Text Available The present study includes an experimental and theoretical investigation of reinforced concrete deep beams containing tensile reinforcement lap splices at constant moment zone under static load. The study included two stages: in the first one, an experimental work included testing of eight simply supported RC deep beams having a total length (L = 2000 mm, overall depth (h= 600 mm and width (b = 150 mm. The tested specimens were divided into three groups to study the effect of main variables: lap length, location of splice, internal confinement (stirrups and external confinement (strengthening by CFRP laminates. The experimental results showed that the use of CFRP as external strengthening in deep beam with lap spliced gives best behavior such as increase in stiffness, decrease in deflection, delaying the cracks appearance and reducing the crack width. The reduction in deflection about (14-21 % than the unstrengthened beam and about (5-14 % than the beam with continuous bars near ultimate load. Also, it was observed that the beams with unstrengthened tensile reinforcement lap splices had three types of cracks: flexural, flexural-shear and splitting cracks while the beams with strengthened tensile reinforcement lap splices or continuous bars don’t observe splitting cracks. In the second stage, a numerical analysis of three dimensional finite element analysis was utilized to explore the behavior of the RC deep beams with tensile reinforcement lap splices, in addition to parametric study of many variables. The comparison between the experimental and theoretical results showed reasonable agreement. The average difference of the deflection at service load was less than 5%.

  20. DeepSpark: A Spark-Based Distributed Deep Learning Framework for Commodity Clusters

    OpenAIRE

    Kim, Hanjoo; Park, Jaehong; Jang, Jaehee; Yoon, Sungroh

    2016-01-01

    The increasing complexity of deep neural networks (DNNs) has made it challenging to exploit existing large-scale data processing pipelines for handling massive data and parameters involved in DNN training. Distributed computing platforms and GPGPU-based acceleration provide a mainstream solution to this computational challenge. In this paper, we propose DeepSpark, a distributed and parallel deep learning framework that exploits Apache Spark on commodity clusters. To support parallel operation...

  1. DANoC: An Efficient Algorithm and Hardware Codesign of Deep Neural Networks on Chip.

    Science.gov (United States)

    Zhou, Xichuan; Li, Shengli; Tang, Fang; Hu, Shengdong; Lin, Zhi; Zhang, Lei

    2017-07-18

    Deep neural networks (NNs) are the state-of-the-art models for understanding the content of images and videos. However, implementing deep NNs in embedded systems is a challenging task, e.g., a typical deep belief network could exhaust gigabytes of memory and result in bandwidth and computational bottlenecks. To address this challenge, this paper presents an algorithm and hardware codesign for efficient deep neural computation. A hardware-oriented deep learning algorithm, named the deep adaptive network, is proposed to explore the sparsity of neural connections. By adaptively removing the majority of neural connections and robustly representing the reserved connections using binary integers, the proposed algorithm could save up to 99.9% memory utility and computational resources without undermining classification accuracy. An efficient sparse-mapping-memory-based hardware architecture is proposed to fully take advantage of the algorithmic optimization. Different from traditional Von Neumann architecture, the deep-adaptive network on chip (DANoC) brings communication and computation in close proximity to avoid power-hungry parameter transfers between on-board memory and on-chip computational units. Experiments over different image classification benchmarks show that the DANoC system achieves competitively high accuracy and efficiency comparing with the state-of-the-art approaches.

  2. Resolution of Singularities Introduced by Hierarchical Structure in Deep Neural Networks.

    Science.gov (United States)

    Nitta, Tohru

    2017-10-01

    We present a theoretical analysis of singular points of artificial deep neural networks, resulting in providing deep neural network models having no critical points introduced by a hierarchical structure. It is considered that such deep neural network models have good nature for gradient-based optimization. First, we show that there exist a large number of critical points introduced by a hierarchical structure in deep neural networks as straight lines, depending on the number of hidden layers and the number of hidden neurons. Second, we derive a sufficient condition for deep neural networks having no critical points introduced by a hierarchical structure, which can be applied to general deep neural networks. It is also shown that the existence of critical points introduced by a hierarchical structure is determined by the rank and the regularity of weight matrices for a specific class of deep neural networks. Finally, two kinds of implementation methods of the sufficient conditions to have no critical points are provided. One is a learning algorithm that can avoid critical points introduced by the hierarchical structure during learning (called avoidant learning algorithm). The other is a neural network that does not have some critical points introduced by the hierarchical structure as an inherent property (called avoidant neural network).

  3. AUC-Maximized Deep Convolutional Neural Fields for Protein Sequence Labeling.

    Science.gov (United States)

    Wang, Sheng; Sun, Siqi; Xu, Jinbo

    2016-09-01

    Deep Convolutional Neural Networks (DCNN) has shown excellent performance in a variety of machine learning tasks. This paper presents Deep Convolutional Neural Fields (DeepCNF), an integration of DCNN with Conditional Random Field (CRF), for sequence labeling with an imbalanced label distribution. The widely-used training methods, such as maximum-likelihood and maximum labelwise accuracy, do not work well on imbalanced data. To handle this, we present a new training algorithm called maximum-AUC for DeepCNF. That is, we train DeepCNF by directly maximizing the empirical Area Under the ROC Curve (AUC), which is an unbiased measurement for imbalanced data. To fulfill this, we formulate AUC in a pairwise ranking framework, approximate it by a polynomial function and then apply a gradient-based procedure to optimize it. Our experimental results confirm that maximum-AUC greatly outperforms the other two training methods on 8-state secondary structure prediction and disorder prediction since their label distributions are highly imbalanced and also has similar performance as the other two training methods on solvent accessibility prediction, which has three equally-distributed labels. Furthermore, our experimental results show that our AUC-trained DeepCNF models greatly outperform existing popular predictors of these three tasks. The data and software related to this paper are available at https://github.com/realbigws/DeepCNF_AUC.

  4. Diabetic retinopathy screening using deep neural network.

    Science.gov (United States)

    Ramachandran, Nishanthan; Hong, Sheng Chiong; Sime, Mary J; Wilson, Graham A

    2017-09-07

    There is a burgeoning interest in the use of deep neural network in diabetic retinal screening. To determine whether a deep neural network could satisfactorily detect diabetic retinopathy that requires referral to an ophthalmologist from a local diabetic retinal screening programme and an international database. Retrospective audit. Diabetic retinal photos from Otago database photographed during October 2016 (485 photos), and 1200 photos from Messidor international database. Receiver operating characteristic curve to illustrate the ability of a deep neural network to identify referable diabetic retinopathy (moderate or worse diabetic retinopathy or exudates within one disc diameter of the fovea). Area under the receiver operating characteristic curve, sensitivity and specificity. For detecting referable diabetic retinopathy, the deep neural network had an area under receiver operating characteristic curve of 0.901 (95% confidence interval 0.807-0.995), with 84.6% sensitivity and 79.7% specificity for Otago and 0.980 (95% confidence interval 0.973-0.986), with 96.0% sensitivity and 90.0% specificity for Messidor. This study has shown that a deep neural network can detect referable diabetic retinopathy with sensitivities and specificities close to or better than 80% from both an international and a domestic (New Zealand) database. We believe that deep neural networks can be integrated into community screening once they can successfully detect both diabetic retinopathy and diabetic macular oedema. © 2017 Royal Australian and New Zealand College of Ophthalmologists.

  5. Deep architectures and deep learning in chemoinformatics: the prediction of aqueous solubility for drug-like molecules.

    Science.gov (United States)

    Lusci, Alessandro; Pollastri, Gianluca; Baldi, Pierre

    2013-07-22

    Shallow machine learning methods have been applied to chemoinformatics problems with some success. As more data becomes available and more complex problems are tackled, deep machine learning methods may also become useful. Here, we present a brief overview of deep learning methods and show in particular how recursive neural network approaches can be applied to the problem of predicting molecular properties. However, molecules are typically described by undirected cyclic graphs, while recursive approaches typically use directed acyclic graphs. Thus, we develop methods to address this discrepancy, essentially by considering an ensemble of recursive neural networks associated with all possible vertex-centered acyclic orientations of the molecular graph. One advantage of this approach is that it relies only minimally on the identification of suitable molecular descriptors because suitable representations are learned automatically from the data. Several variants of this approach are applied to the problem of predicting aqueous solubility and tested on four benchmark data sets. Experimental results show that the performance of the deep learning methods matches or exceeds the performance of other state-of-the-art methods according to several evaluation metrics and expose the fundamental limitations arising from training sets that are too small or too noisy. A Web-based predictor, AquaSol, is available online through the ChemDB portal ( cdb.ics.uci.edu ) together with additional material.

  6. Some Challenges of Deep Mining†

    Directory of Open Access Journals (Sweden)

    Charles Fairhurst

    2017-08-01

    Full Text Available An increased global supply of minerals is essential to meet the needs and expectations of a rapidly rising world population. This implies extraction from greater depths. Autonomous mining systems, developed through sustained R&D by equipment suppliers, reduce miner exposure to hostile work environments and increase safety. This places increased focus on “ground control” and on rock mechanics to define the depth to which minerals may be extracted economically. Although significant efforts have been made since the end of World War II to apply mechanics to mine design, there have been both technological and organizational obstacles. Rock in situ is a more complex engineering material than is typically encountered in most other engineering disciplines. Mining engineering has relied heavily on empirical procedures in design for thousands of years. These are no longer adequate to address the challenges of the 21st century, as mines venture to increasingly greater depths. The development of the synthetic rock mass (SRM in 2008 provides researchers with the ability to analyze the deformational behavior of rock masses that are anisotropic and discontinuous—attributes that were described as the defining characteristics of in situ rock by Leopold Müller, the president and founder of the International Society for Rock Mechanics (ISRM, in 1966. Recent developments in the numerical modeling of large-scale mining operations (e.g., caving using the SRM reveal unanticipated deformational behavior of the rock. The application of massive parallelization and cloud computational techniques offers major opportunities: for example, to assess uncertainties in numerical predictions; to establish the mechanics basis for the empirical rules now used in rock engineering and their validity for the prediction of rock mass behavior beyond current experience; and to use the discrete element method (DEM in the optimization of deep mine design. For the first time, mining

  7. DeepInfer: open-source deep learning deployment toolkit for image-guided therapy

    Science.gov (United States)

    Mehrtash, Alireza; Pesteie, Mehran; Hetherington, Jorden; Behringer, Peter A.; Kapur, Tina; Wells, William M.; Rohling, Robert; Fedorov, Andriy; Abolmaesumi, Purang

    2017-03-01

    Deep learning models have outperformed some of the previous state-of-the-art approaches in medical image analysis. Instead of using hand-engineered features, deep models attempt to automatically extract hierarchical representations at multiple levels of abstraction from the data. Therefore, deep models are usually considered to be more flexible and robust solutions for image analysis problems compared to conventional computer vision models. They have demonstrated significant improvements in computer-aided diagnosis and automatic medical image analysis applied to such tasks as image segmentation, classification and registration. However, deploying deep learning models often has a steep learning curve and requires detailed knowledge of various software packages. Thus, many deep models have not been integrated into the clinical research work ows causing a gap between the state-of-the-art machine learning in medical applications and evaluation in clinical research procedures. In this paper, we propose "DeepInfer" - an open-source toolkit for developing and deploying deep learning models within the 3D Slicer medical image analysis platform. Utilizing a repository of task-specific models, DeepInfer allows clinical researchers and biomedical engineers to deploy a trained model selected from the public registry, and apply it to new data without the need for software development or configuration. As two practical use cases, we demonstrate the application of DeepInfer in prostate segmentation for targeted MRI-guided biopsy and identification of the target plane in 3D ultrasound for spinal injections.

  8. Building Watson: An Overview of the DeepQA Project

    OpenAIRE

    Ferrucci, David; Brown, Eric; Chu-Carroll, Jennifer; Fan, James; Gondek, David; Kalyanpur, Aditya A.; Lally, Adam; Murdock, J. William; Nyberg, Eric; Prager, John; Schlaefer, Nico; Welty, Chris

    2010-01-01

    IBM Research undertook a challenge to build a computer system that could compete at the human champion level in real time on the American TV Quiz show, Jeopardy! The extent of the challenge includes fielding a real-time automatic contestant on the show, not merely a laboratory exercise. The Jeopardy! Challenge helped us address requirements that led to the design of the DeepQA architecture and the implementation of Watson. After 3 years of intense research and development by a core team of ab...

  9. Deep and surface learning in problem-based learning: a review of the literature.

    Science.gov (United States)

    Dolmans, Diana H J M; Loyens, Sofie M M; Marcq, Hélène; Gijbels, David

    2016-12-01

    In problem-based learning (PBL), implemented worldwide, students learn by discussing professionally relevant problems enhancing application and integration of knowledge, which is assumed to encourage students towards a deep learning approach in which students are intrinsically interested and try to understand what is being studied. This review investigates: (1) the effects of PBL on students' deep and surface approaches to learning, (2) whether and why these effects do differ across (a) the context of the learning environment (single vs. curriculum wide implementation), and (b) study quality. Studies were searched dealing with PBL and students' approaches to learning. Twenty-one studies were included. The results indicate that PBL does enhance deep learning with a small positive average effect size of .11 and a positive effect in eleven of the 21 studies. Four studies show a decrease in deep learning and six studies show no effect. PBL does not seem to have an effect on surface learning as indicated by a very small average effect size (.08) and eleven studies showing no increase in the surface approach. Six studies demonstrate a decrease and four an increase in surface learning. It is concluded that PBL does seem to enhance deep learning and has little effect on surface learning, although more longitudinal research using high quality measurement instruments is needed to support this conclusion with stronger evidence. Differences cannot be explained by the study quality but a curriculum wide implementation of PBL has a more positive impact on the deep approach (effect size .18) compared to an implementation within a single course (effect size of -.05). PBL is assumed to enhance active learning and students' intrinsic motivation, which enhances deep learning. A high perceived workload and assessment that is perceived as not rewarding deep learning are assumed to enhance surface learning.

  10. Lyman Break Galaxies in the Hubble Ultra Deep Field through Deep U-Band Imaging

    Science.gov (United States)

    Rafelski, Marc; Wolfe, A. M.; Cooke, J.; Chen, H. W.; Armandroff, T. E.; Wirth, G. D.

    2009-12-01

    We introduce an extremely deep U-band image taken of the Hubble Ultra Deep Field (HUDF), with a one sigma depth of 30.7 mag arcsec-2 and a detection limiting magnitude of 28 mag arcsec-2. The observations were carried out on the Keck I telescope using the LRIS-B detector. The U-band image substantially improves the accuracy of photometric redshift measurements of faint galaxies in the HUDF at z=[2.5,3.5]. The U-band for these galaxies is attenuated by lyman limit absorption, allowing for more reliable selections of candidate Lyman Break Galaxies (LBGs) than from photometric redshifts without U-band. We present a reliable sample of 300 LBGs at z=[2.5,3.5] in the HUDF. Accurate redshifts of faint galaxies at z=[2.5,3.5] are needed to obtain empirical constraints on the star formation efficiency of neutral gas at high redshift. Wolfe & Chen (2006) showed that the star formation rate (SFR) density in damped Ly-alpha absorption systems (DLAs) at z=[2.5,3.5] is significantly lower than predicted by the Kennicutt-Schmidt law for nearby galaxies. One caveat to this result that we wish to test is whether LBGs are embedded in DLAs. If in-situ star formation is occurring in DLAs, we would see it as extended low surface brightness emission around LBGs. We shall use the more accurate photometric redshifts to create a sample of LBGs around which we will look for extended emission in the more sensitive and higher resolution HUDF images. The absence of extended emission would put limits on the SFR density of DLAs associated with LBGs at high redshift. On the other hand, detection of faint emission on scales large compared to the bright LBG cores would indicate the presence of in situ star formation in those DLAs. Such gas would presumably fuel the higher star formation rates present in the LBG cores.

  11. Ultra Deep Wave Equation Imaging and Illumination

    Energy Technology Data Exchange (ETDEWEB)

    Alexander M. Popovici; Sergey Fomel; Paul Sava; Sean Crawley; Yining Li; Cristian Lupascu

    2006-09-30

    In this project we developed and tested a novel technology, designed to enhance seismic resolution and imaging of ultra-deep complex geologic structures by using state-of-the-art wave-equation depth migration and wave-equation velocity model building technology for deeper data penetration and recovery, steeper dip and ultra-deep structure imaging, accurate velocity estimation for imaging and pore pressure prediction and accurate illumination and amplitude processing for extending the AVO prediction window. Ultra-deep wave-equation imaging provides greater resolution and accuracy under complex geologic structures where energy multipathing occurs, than what can be accomplished today with standard imaging technology. The objective of the research effort was to examine the feasibility of imaging ultra-deep structures onshore and offshore, by using (1) wave-equation migration, (2) angle-gathers velocity model building, and (3) wave-equation illumination and amplitude compensation. The effort consisted of answering critical technical questions that determine the feasibility of the proposed methodology, testing the theory on synthetic data, and finally applying the technology for imaging ultra-deep real data. Some of the questions answered by this research addressed: (1) the handling of true amplitudes in the downward continuation and imaging algorithm and the preservation of the amplitude with offset or amplitude with angle information required for AVO studies, (2) the effect of several imaging conditions on amplitudes, (3) non-elastic attenuation and approaches for recovering the amplitude and frequency, (4) the effect of aperture and illumination on imaging steep dips and on discriminating the velocities in the ultra-deep structures. All these effects were incorporated in the final imaging step of a real data set acquired specifically to address ultra-deep imaging issues, with large offsets (12,500 m) and long recording time (20 s).

  12. Hydro-mechanical deep drawing of rolled magnesium sheets

    Energy Technology Data Exchange (ETDEWEB)

    Bach, F.W.; Rodman, M.; Rossberg, A. [Hannover Univ., Garbsen (Germany). Inst. of Materials Science; Behrens, B.A.; Vogt, O. [Hannover Univ., Garbsen (DE). Inst. of Metal Forming and Metal Forming Machine Tools (IFUM)

    2005-12-01

    Magnesium sheets offer high specific properties which make them very attractive in modern light weight constructions. The main obstacles for a wider usage are their high production costs, the poor corrosion properties and the limited ductility. Until today, forming processes have to be conducted at temperatures well above T=220 C. In the first place, this is a cost factor. Moreover, technical aspects, such as grain growth or the limited use of lubrication speak against high temperatures. The first aim of the presented research work is to increase the ductility at lower temperatures by alloy modification and by an adapted rolling technology. The key factor to reach isotropic mechanical properties and increased limit drawing ratios in deep drawing tools, is to achieve fine, homogeneous microstructures. This can be done by cross rolling at moderate temperatures. The heat treatment has to be adapted accordingly. In a second stage, hydro-mechanical deep drawing experiments were carried out at elevated temperature. The results show that the forming behaviour of the tested Mg-alloys is considerably improved compared to conventional deep drawing. (orig.)

  13. Factors governing the deep ventilation of the Red Sea

    KAUST Repository

    Papadopoulos, Vassilis P.

    2015-11-19

    A variety of data based on hydrographic measurements, satellite observations, reanalysis databases, and meteorological observations are used to explore the interannual variability and factors governing the deep water formation in the northern Red Sea. Historical and recent hydrographic data consistently indicate that the ventilation of the near-bottom layer in the Red Sea is a robust feature of the thermohaline circulation. Dense water capable to reach the bottom layers of the Red Sea can be regularly produced mostly inside the Gulfs of Aqaba and Suez. Occasionally, during colder than usual winters, deep water formation may also take place over coastal areas in the northernmost end of the open Red Sea just outside the Gulfs of Aqaba and Suez. However, the origin as well as the amount of deep waters exhibit considerable interannual variability depending not only on atmospheric forcing but also on the water circulation over the northern Red Sea. Analysis of several recent winters shows that the strength of the cyclonic gyre prevailing in the northernmost part of the basin can effectively influence the sea surface temperature (SST) and intensify or moderate the winter surface cooling. Upwelling associated with periods of persistent gyre circulation lowers the SST over the northernmost part of the Red Sea and can produce colder than normal winter SST even without extreme heat loss by the sea surface. In addition, the occasional persistence of the cyclonic gyre feeds the surface layers of the northern Red Sea with nutrients, considerably increasing the phytoplankton biomass.

  14. The Next Era: Deep Learning in Pharmaceutical Research.

    Science.gov (United States)

    Ekins, Sean

    2016-11-01

    Over the past decade we have witnessed the increasing sophistication of machine learning algorithms applied in daily use from internet searches, voice recognition, social network software to machine vision software in cameras, phones, robots and self-driving cars. Pharmaceutical research has also seen its fair share of machine learning developments. For example, applying such methods to mine the growing datasets that are created in drug discovery not only enables us to learn from the past but to predict a molecule's properties and behavior in future. The latest machine learning algorithm garnering significant attention is deep learning, which is an artificial neural network with multiple hidden layers. Publications over the last 3 years suggest that this algorithm may have advantages over previous machine learning methods and offer a slight but discernable edge in predictive performance. The time has come for a balanced review of this technique but also to apply machine learning methods such as deep learning across a wider array of endpoints relevant to pharmaceutical research for which the datasets are growing such as physicochemical property prediction, formulation prediction, absorption, distribution, metabolism, excretion and toxicity (ADME/Tox), target prediction and skin permeation, etc. We also show that there are many potential applications of deep learning beyond cheminformatics. It will be important to perform prospective testing (which has been carried out rarely to date) in order to convince skeptics that there will be benefits from investing in this technique.

  15. Admittance spectroscopy or deep level transient spectroscopy: A contrasting juxtaposition

    Science.gov (United States)

    Bollmann, Joachim; Venter, Andre

    2018-04-01

    A comprehensive understanding of defects in semiconductors remains of primary importance. In this paper the effectiveness of two of the most commonly used semiconductor defect spectroscopy techniques, viz. deep level transient spectroscopy (DLTS) and admittance spectroscopy (AS) are reviewed. The analysis of defects present in commercially available SiC diodes shows that admittance spectroscopy allows the identification of deep traps with reduced measurement effort compared to deep Level Transient Spectroscopy (DLTS). Besides the N-donor, well-studied intrinsic defects were detected in these diodes. Determination of their activation energy and defect density, using the two techniques, confirm that the sensitivity of AS is comparable to that of DLTS while, due to its well defined peak shape, the spectroscopic resolution is superior. Additionally, admittance spectroscopy can analyze faster emission processes which make the study of shallow defects more practical and even that of shallow dopant levels, possible. A comparative summary for the relevant spectroscopic features of the two capacitance methods are presented.

  16. Efficient collective swimming by harnessing vortices through deep reinforcement learning.

    Science.gov (United States)

    Verma, Siddhartha; Novati, Guido; Koumoutsakos, Petros

    2018-06-05

    Fish in schooling formations navigate complex flow fields replete with mechanical energy in the vortex wakes of their companions. Their schooling behavior has been associated with evolutionary advantages including energy savings, yet the underlying physical mechanisms remain unknown. We show that fish can improve their sustained propulsive efficiency by placing themselves in appropriate locations in the wake of other swimmers and intercepting judiciously their shed vortices. This swimming strategy leads to collective energy savings and is revealed through a combination of high-fidelity flow simulations with a deep reinforcement learning (RL) algorithm. The RL algorithm relies on a policy defined by deep, recurrent neural nets, with long-short-term memory cells, that are essential for capturing the unsteadiness of the two-way interactions between the fish and the vortical flow field. Surprisingly, we find that swimming in-line with a leader is not associated with energetic benefits for the follower. Instead, "smart swimmer(s)" place themselves at off-center positions, with respect to the axis of the leader(s) and deform their body to synchronize with the momentum of the oncoming vortices, thus enhancing their swimming efficiency at no cost to the leader(s). The results confirm that fish may harvest energy deposited in vortices and support the conjecture that swimming in formation is energetically advantageous. Moreover, this study demonstrates that deep RL can produce navigation algorithms for complex unsteady and vortical flow fields, with promising implications for energy savings in autonomous robotic swarms.

  17. Skin Lesion Analysis towards Melanoma Detection Using Deep Learning Network

    Directory of Open Access Journals (Sweden)

    Yuexiang Li

    2018-02-01

    Full Text Available Skin lesions are a severe disease globally. Early detection of melanoma in dermoscopy images significantly increases the survival rate. However, the accurate recognition of melanoma is extremely challenging due to the following reasons: low contrast between lesions and skin, visual similarity between melanoma and non-melanoma lesions, etc. Hence, reliable automatic detection of skin tumors is very useful to increase the accuracy and efficiency of pathologists. In this paper, we proposed two deep learning methods to address three main tasks emerging in the area of skin lesion image processing, i.e., lesion segmentation (task 1, lesion dermoscopic feature extraction (task 2 and lesion classification (task 3. A deep learning framework consisting of two fully convolutional residual networks (FCRN is proposed to simultaneously produce the segmentation result and the coarse classification result. A lesion index calculation unit (LICU is developed to refine the coarse classification results by calculating the distance heat-map. A straight-forward CNN is proposed for the dermoscopic feature extraction task. The proposed deep learning frameworks were evaluated on the ISIC 2017 dataset. Experimental results show the promising accuracies of our frameworks, i.e., 0.753 for task 1, 0.848 for task 2 and 0.912 for task 3 were achieved.

  18. Skin Lesion Analysis towards Melanoma Detection Using Deep Learning Network.

    Science.gov (United States)

    Li, Yuexiang; Shen, Linlin

    2018-02-11

    Skin lesions are a severe disease globally. Early detection of melanoma in dermoscopy images significantly increases the survival rate. However, the accurate recognition of melanoma is extremely challenging due to the following reasons: low contrast between lesions and skin, visual similarity between melanoma and non-melanoma lesions, etc. Hence, reliable automatic detection of skin tumors is very useful to increase the accuracy and efficiency of pathologists. In this paper, we proposed two deep learning methods to address three main tasks emerging in the area of skin lesion image processing, i.e., lesion segmentation (task 1), lesion dermoscopic feature extraction (task 2) and lesion classification (task 3). A deep learning framework consisting of two fully convolutional residual networks (FCRN) is proposed to simultaneously produce the segmentation result and the coarse classification result. A lesion index calculation unit (LICU) is developed to refine the coarse classification results by calculating the distance heat-map. A straight-forward CNN is proposed for the dermoscopic feature extraction task. The proposed deep learning frameworks were evaluated on the ISIC 2017 dataset. Experimental results show the promising accuracies of our frameworks, i.e., 0.753 for task 1, 0.848 for task 2 and 0.912 for task 3 were achieved.

  19. Skin Lesion Analysis towards Melanoma Detection Using Deep Learning Network

    Science.gov (United States)

    2018-01-01

    Skin lesions are a severe disease globally. Early detection of melanoma in dermoscopy images significantly increases the survival rate. However, the accurate recognition of melanoma is extremely challenging due to the following reasons: low contrast between lesions and skin, visual similarity between melanoma and non-melanoma lesions, etc. Hence, reliable automatic detection of skin tumors is very useful to increase the accuracy and efficiency of pathologists. In this paper, we proposed two deep learning methods to address three main tasks emerging in the area of skin lesion image processing, i.e., lesion segmentation (task 1), lesion dermoscopic feature extraction (task 2) and lesion classification (task 3). A deep learning framework consisting of two fully convolutional residual networks (FCRN) is proposed to simultaneously produce the segmentation result and the coarse classification result. A lesion index calculation unit (LICU) is developed to refine the coarse classification results by calculating the distance heat-map. A straight-forward CNN is proposed for the dermoscopic feature extraction task. The proposed deep learning frameworks were evaluated on the ISIC 2017 dataset. Experimental results show the promising accuracies of our frameworks, i.e., 0.753 for task 1, 0.848 for task 2 and 0.912 for task 3 were achieved. PMID:29439500

  20. Study of microorganisms present in deep geologic formations

    International Nuclear Information System (INIS)

    Camus, H.; Lion, R.; Bianchi, A.; Garcin, J.

    1987-01-01

    This work has been executed in the scope of the studies on high activity radioactive wastes storage in deep geological environments. The authors make reference to an as complete as possible literature on the existence of microorganisms in those environments or under similar conditions. Then they describe the equipment and methods they have implemented to perform their study of the populations present in three deep-reaching drill-holes in Auriat (France), Mol (Belgique) and Troon (Great Britain). The results of the study exhibit the presence of a certain biological activity, well adapted to that particular life environment. Strains appear to be very varied from the taxonomic point of view and seemingly show an important potential of mineral alteration when provided with an adequate source of energy. Complementary studies, using advanced techniques such as those employed during the work forming the basis of this paper, seem necessary for a more accurate evaluation of long-term risks of perturbation of a deep storage site [fr

  1. Prediction of visual saliency in video with deep CNNs

    Science.gov (United States)

    Chaabouni, Souad; Benois-Pineau, Jenny; Hadar, Ofer

    2016-09-01

    Prediction of visual saliency in images and video is a highly researched topic. Target applications include Quality assessment of multimedia services in mobile context, video compression techniques, recognition of objects in video streams, etc. In the framework of mobile and egocentric perspectives, visual saliency models cannot be founded only on bottom-up features, as suggested by feature integration theory. The central bias hypothesis, is not respected neither. In this case, the top-down component of human visual attention becomes prevalent. Visual saliency can be predicted on the basis of seen data. Deep Convolutional Neural Networks (CNN) have proven to be a powerful tool for prediction of salient areas in stills. In our work we also focus on sensitivity of human visual system to residual motion in a video. A Deep CNN architecture is designed, where we incorporate input primary maps as color values of pixels and magnitude of local residual motion. Complementary contrast maps allow for a slight increase of accuracy compared to the use of color and residual motion only. The experiments show that the choice of the input features for the Deep CNN depends on visual task:for th eintersts in dynamic content, the 4K model with residual motion is more efficient, and for object recognition in egocentric video the pure spatial input is more appropriate.

  2. Integrating shallow and deep knowledge in the design of an on-line process monitoring system

    Energy Technology Data Exchange (ETDEWEB)

    Gallanti, M.; Gilardoni, L.; Guida, G.; Stefanini, A.; Tomada, L.

    1989-01-01

    Monitoring and malfunctions diagnosis of complex industrial plants involves, in addition to shallow empirical knowledge about plant operation, also deep knowledge about structure and function. This paper presents the results obtained in the design and experimentation of PROP and PROP-2 systems, devoted to on-line monitoring and diagnosis of pollution phenomena in the cycle water of a thermal power plant. In particular, it focuses on PROP-2 architecture, with encompasses a four-level hierarchical knowledge base including both empirical knowledge and a deep model of the plant. Shallow knowledge is represented by production rules and event-graphs (a formalism for expressing procedural knowledge), while deep knowledge is expressed using a representation language based on the concept of component. One major contribution of the proposed approach has been to show in a running experimental system that a suitable blend of shallow and deep knowledge can offer substantial advantages over a single paradigm.

  3. Equivalence between deep energy-dependent and shallow angular momentum dependent potentials

    International Nuclear Information System (INIS)

    Fiedeldey, H.; Sofianos, S.A.; Papastylianos, A.; Amos, K.A.; Allen, L.J.

    1989-01-01

    Recently Baye showed that supersymmetry can be applied to determine a shallow l-dependent potential phase equivalent to a deep potential, assumed to be energy-independent and have Panli forbidden states (PFS), for α-α scattering. The PFS are eliminated by this procedure. Such deep potentials are generated as equivalent local potentials (ELP) to the Resonating Group Model (RGM) and are generally energy-dependent. To eliminate this E-dependence as required for the application of Baye's method, l-dependent, but E-independent, deep local potentials were generated by the exact inversion method of Marchenko. Subsequently, the supersymmetric method was used to eliminate the PFS, ensuring that the generalized Levinson theorem is satisfied. As an example, the method was applied to the simple model of two dineutrons scattering in the RGM, where the deep ELP of Horiuchi has a substantial energy-dependence and one PFS only for l=O. 16 refs., 5 figs

  4. A Review of Deep Learning Methods and Applications for Unmanned Aerial Vehicles

    Directory of Open Access Journals (Sweden)

    Adrian Carrio

    2017-01-01

    Full Text Available Deep learning is recently showing outstanding results for solving a wide variety of robotic tasks in the areas of perception, planning, localization, and control. Its excellent capabilities for learning representations from the complex data acquired in real environments make it extremely suitable for many kinds of autonomous robotic applications. In parallel, Unmanned Aerial Vehicles (UAVs are currently being extensively applied for several types of civilian tasks in applications going from security, surveillance, and disaster rescue to parcel delivery or warehouse management. In this paper, a thorough review has been performed on recent reported uses and applications of deep learning for UAVs, including the most relevant developments as well as their performances and limitations. In addition, a detailed explanation of the main deep learning techniques is provided. We conclude with a description of the main challenges for the application of deep learning for UAV-based solutions.

  5. Active semi-supervised learning method with hybrid deep belief networks.

    Science.gov (United States)

    Zhou, Shusen; Chen, Qingcai; Wang, Xiaolong

    2014-01-01

    In this paper, we develop a novel semi-supervised learning algorithm called active hybrid deep belief networks (AHD), to address the semi-supervised sentiment classification problem with deep learning. First, we construct the previous several hidden layers using restricted Boltzmann machines (RBM), which can reduce the dimension and abstract the information of the reviews quickly. Second, we construct the following hidden layers using convolutional restricted Boltzmann machines (CRBM), which can abstract the information of reviews effectively. Third, the constructed deep architecture is fine-tuned by gradient-descent based supervised learning with an exponential loss function. Finally, active learning method is combined based on the proposed deep architecture. We did several experiments on five sentiment classification datasets, and show that AHD is competitive with previous semi-supervised learning algorithm. Experiments are also conducted to verify the effectiveness of our proposed method with different number of labeled reviews and unlabeled reviews respectively.

  6. Xe isotopic constraints on cycling of deep Earth volatiles

    Science.gov (United States)

    Parai, R.; Mukhopadhyay, S.

    2017-12-01

    The modern deep Earth volatile budget reflects primordial volatiles delivered during accretion, radiogenic ingrowth of volatile species (e.g., 40Ar produced by 40K decay), outgassing in association with mantle processing, and regassing via subduction. The noble gases are unique volatile tracers in that they are chemically inert, but are thought to be trapped within hydrous alteration phases in downwelling lithologies. Noble gases thus provide a tracer of volatile transport between the deep Earth and surface reservoirs. Constraints on the fluxes of noble gases between deep Earth and surface reservoirs over time can accordingly be used to provide insight into temperature conditions at subduction zones, limits on volatile cycling, and the evolving distribution of major volatile species in terrestrial reservoirs over time. Xe isotope systematics in mantle-derived rocks show that 80-90% of the mantle Xe budget is derived from recycling of atmospheric Xe, indicating that atmospheric Xe is retained in subducting slabs beyond depths of magma generation in subduction zones over Earth history. We present an integrated model of Xe cycling between the mantle and atmosphere in association with mantle processing over Earth history. We test a wide variety of outgassing and regassing rates and take the evolution of the atmospheric Xe isotopic composition [e.g., 1] into account. Models in which the deep Earth transitions from a net outgassing to net regassing regime best satisfy Xe isotopic constraints from mantle-derived rocks [2-6]. [1] Avice et al., 2017; Nature Communications, 8; [2] Mukhopadhyay, 2012, Nature 486, 101-104; [3] Parai et al., 2012, EPSL 359-360, 227-239; [4] Parai and Mukhopadhay, 2015, G-cubed 16, 719-735; [5] Peto et al., 2013, EPSL 369-370, 13-23; [6] Tucker et al., 2012, EPSL 355-356, 244-254.

  7. A deep learning approach for fetal QRS complex detection.

    Science.gov (United States)

    Zhong, Wei; Liao, Lijuan; Guo, Xuemei; Wang, Guoli

    2018-04-20

    Non-invasive foetal electrocardiography (NI-FECG) has the potential to provide more additional clinical information for detecting and diagnosing fetal diseases. We propose and demonstrate a deep learning approach for fetal QRS complex detection from raw NI-FECG signals by using a convolutional neural network (CNN) model. The main objective is to investigate whether reliable fetal QRS complex detection performance can still be obtained from features of single-channel NI-FECG signals, without canceling maternal ECG (MECG) signals. A deep learning method is proposed for recognizing fetal QRS complexes. Firstly, we collect data from set-a of the PhysioNet/computing in Cardiology Challenge database. The sample entropy method is used for signal quality assessment. Part of the bad quality signals is excluded in the further analysis. Secondly, in the proposed method, the features of raw NI-FECG signals are normalized before they are fed to a CNN classifier to perform fetal QRS complex detection. We use precision, recall, F-measure and accuracy as the evaluation metrics to assess the performance of fetal QRS complex detection. The proposed deep learning method can achieve relatively high precision (75.33%), recall (80.54%), and F-measure scores (77.85%) compared with three other well-known pattern classification methods, namely KNN, naive Bayes and SVM. the proposed deep learning method can attain reliable fetal QRS complex detection performance from the raw NI-FECG signals without canceling MECG signals. In addition, the influence of different activation functions and signal quality assessment on classification performance are evaluated, and results show that Relu outperforms the Sigmoid and Tanh on this particular task, and better classification performance is obtained with the signal quality assessment step in this study.

  8. FY-2015 Methyl Iodide Deep-Bed Adsorption Test Report

    Energy Technology Data Exchange (ETDEWEB)

    Soelberg, Nicholas Ray [Idaho National Lab. (INL), Idaho Falls, ID (United States); Watson, Tony Leroy [Idaho National Lab. (INL), Idaho Falls, ID (United States)

    2015-09-30

    Nuclear fission produces fission and activation products, including iodine-129, which could evolve into used fuel reprocessing facility off-gas systems, and could require off-gas control to limit air emissions to levels within acceptable emission limits. Deep-bed methyl iodide adsorption testing has continued in Fiscal Year 2015 according to a multi-laboratory methyl iodide adsorption test plan. Updates to the deep-bed test system have also been performed to enable the inclusion of evaporated HNO3 and increased NO2 concentrations in future tests. This report summarizes the result of those activities. Test results showed that iodine adsorption from gaseous methyl iodide using reduced silver zeolite (AgZ) resulted in initial iodine decontamination factors (DFs, ratios of uncontrolled and controlled total iodine levels) under 1,000 for the conditions of the long-duration test performed this year (45 ppm CH3I, 1,000 ppm each NO and NO2, very low H2O levels [3 ppm] in balance air). The mass transfer zone depth exceeded the cumulative 5-inch depth of 4 bed segments, which is deeper than the 2-4 inch depth estimated for the mass transfer zone for adsorbing I2 using AgZ in prior deep-bed tests. The maximum iodine adsorption capacity for the AgZ under the conditions of this test was 6.2% (6.2 g adsorbed I per 100 g sorbent). The maximum Ag utilization was 51%. Additional deep-bed testing and analyses are recommended to (a) expand the data base for methyl iodide adsorption and (b) provide more data for evaluating organic iodide reactions and reaction byproducts for different potential adsorption conditions.

  9. FY-2015 Methyl Iodide Deep-Bed Adsorption Test Report

    International Nuclear Information System (INIS)

    Soelberg, Nicholas Ray; Watson, Tony Leroy

    2015-01-01

    Nuclear fission produces fission and activation products, including iodine-129, which could evolve into used fuel reprocessing facility off-gas systems, and could require off-gas control to limit air emissions to levels within acceptable emission limits. Deep-bed methyl iodide adsorption testing has continued in Fiscal Year 2015 according to a multi-laboratory methyl iodide adsorption test plan. Updates to the deep-bed test system have also been performed to enable the inclusion of evaporated HNO 3 and increased NO 2 concentrations in future tests. This report summarizes the result of those activities. Test results showed that iodine adsorption from gaseous methyl iodide using reduced silver zeolite (AgZ) resulted in initial iodine decontamination factors (DFs, ratios of uncontrolled and controlled total iodine levels) under 1,000 for the conditions of the long-duration test performed this year (45 ppm CH3I, 1,000 ppm each NO and NO 2 , very low H 2 O levels [3 ppm] in balance air). The mass transfer zone depth exceeded the cumulative 5-inch depth of 4 bed segments, which is deeper than the 2-4 inch depth estimated for the mass transfer zone for adsorbing I 2 using AgZ in prior deep-bed tests. The maximum iodine adsorption capacity for the AgZ under the conditions of this test was 6.2% (6.2 g adsorbed I per 100 g sorbent). The maximum Ag utilization was 51%. Additional deep-bed testing and analyses are recommended to (a) expand the data base for methyl iodide adsorption and (b) provide more data for evaluating organic iodide reactions and reaction byproducts for different potential adsorption conditions.

  10. Numerical simulation of phenomenon on zonal disintegration in deep underground mining in case of unsupported roadway

    Science.gov (United States)

    Han, Fengshan; Wu, Xinli; Li, Xia; Zhu, Dekang

    2018-02-01

    Zonal disintegration phenomenon was found in deep mining roadway surrounding rock. It seriously affects the safety of mining and underground engineering and it may lead to the occurrence of natural disasters. in deep mining roadway surrounding rock, tectonic stress in deep mining roadway rock mass, horizontal stress is much greater than the vertical stress, When the direction of maximum principal stress is parallel to the axis of the roadway in deep mining, this is the main reasons for Zonal disintegration phenomenon. Using ABAQUS software to numerical simulation of the three-dimensional model of roadway rupture formation process systematically, and the study shows that when The Direction of maximum main stress in deep underground mining is along the roadway axial direction, Zonal disintegration phenomenon in deep underground mining is successfully reproduced by our numerical simulation..numerical simulation shows that using ABAQUA simulation can reproduce Zonal disintegration phenomenon and the formation process of damage of surrounding rock can be reproduced. which have important engineering practical significance.

  11. Species-energy relationship in the deep sea: A test using the Quaternary fossil record

    Science.gov (United States)

    Hunt, G.; Cronin, T. M.; Roy, K.

    2005-01-01

    Little is known about the processes regulating species richness in deep-sea communities. Here we take advantage of natural experiments involving climate change to test whether predictions of the species-energy hypothesis hold in the deep sea. In addition, we test for the relationship between temperature and species richness predicted by a recent model based on biochemical kinetics of metabolism. Using the deep-sea fossil record of benthic foraminifera and statistical meta-analyses of temperature-richness and productivity-richness relationships in 10 deep-sea cores, we show that temperature but not productivity is a significant predictor of species richness over the past c. 130 000 years. Our results not only show that the temperature-richness relationship in the deep-sea is remarkably similar to that found in terrestrial and shallow marine habitats, but also that species richness tracks temperature change over geological time, at least on scales of c. 100 000 years. Thus, predicting biotic response to global climate change in the deep sea would require better understanding of how temperature regulates the occurrences and geographical ranges of species. ??2005 Blackwell Publishing Ltd/CNRS.

  12. DeepPVP: phenotype-based prioritization of causative variants using deep learning

    KAUST Repository

    Boudellioua, Imene; Kulmanov, Maxat; Schofield, Paul N; Gkoutos, Georgios V; Hoehndorf, Robert

    2018-01-01

    phenotype-based methods that use similar features. DeepPVP is freely available at https://github.com/bio-ontology-research-group/phenomenet-vp Conclusions: DeepPVP further improves on existing variant prioritization methods both in terms of speed as well

  13. Assessment of deep geological environment condition

    International Nuclear Information System (INIS)

    Bae, Dae Seok; Han, Kyung Won; Joen, Kwan Sik

    2003-05-01

    The main tasks of geoscientific study in the 2nd stage was characterized focusing mainly on a near-field condition of deep geologic environment, and aimed to generate the geologic input data for a Korean reference disposal system for high level radioactive wastes and to establish site characterization methodology, including neotectonic features, fracture systems and mechanical properties of plutonic rocks, and hydrogeochemical characteristics. The preliminary assessment of neotectonics in the Korean peninsula was performed on the basis of seismicity recorded, Quarternary faults investigated, uplift characteristics studied on limited areas, distribution of the major regional faults and their characteristics. The local fracture system was studied in detail from the data obtained from deep boreholes in granitic terrain. Through this deep drilling project, the geometrical and hydraulic properties of different fracture sets are statistically analysed on a block scale. The mechanical properties of intact rocks were evaluated from the core samples by laboratory testing and the in-situ stress conditions were estimated by a hydro fracturing test in the boreholes. The hydrogeochemical conditions in the deep boreholes were characterized based on hydrochemical composition and isotopic signatures and were attempted to assess the interrelation with a major fracture system. The residence time of deep groundwater was estimated by C-14 dating. For the travel time of groundwater between the boreholes, the methodology and equipment for tracer test were established

  14. Molecular analysis of deep subsurface bacteria

    International Nuclear Information System (INIS)

    Jimenez Baez, L.E.

    1989-09-01

    Deep sediments samples from site C10a, in Appleton, and sites, P24, P28, and P29, at the Savannah River Site (SRS), near Aiken, South Carolina were studied to determine their microbial community composition, DNA homology and mol %G+C. Different geological formations with great variability in hydrogeological parameters were found across the depth profile. Phenotypic identification of deep subsurface bacteria underestimated the bacterial diversity at the three SRS sites, since bacteria with the same phenotype have different DNA composition and less than 70% DNA homology. Total DNA hybridization and mol %G+C analysis of deep sediment bacterial isolates suggested that each formation is comprised of different microbial communities. Depositional environment was more important than site and geological formation on the DNA relatedness between deep subsurface bacteria, since more 70% of bacteria with 20% or more of DNA homology came from the same depositional environments. Based on phenotypic and genotypic tests Pseudomonas spp. and Acinetobacter spp.-like bacteria were identified in 85 million years old sediments. This suggests that these microbial communities might have been adapted during a long period of time to the environmental conditions of the deep subsurface

  15. Preface: Deep Slab and Mantle Dynamics

    Science.gov (United States)

    Suetsugu, Daisuke; Bina, Craig R.; Inoue, Toru; Wiens, Douglas A.

    2010-11-01

    We are pleased to publish this special issue of the journal Physics of the Earth and Planetary Interiors entitled "Deep Slab and Mantle Dynamics". This issue is an outgrowth of the international symposium "Deep Slab and Mantle Dynamics", which was held on February 25-27, 2009, in Kyoto, Japan. This symposium was organized by the "Stagnant Slab Project" (SSP) research group to present the results of the 5-year project and to facilitate intensive discussion with well-known international researchers in related fields. The SSP and the symposium were supported by a Grant-in-Aid for Scientific Research (16075101) from the Ministry of Education, Culture, Sports, Science and Technology of the Japanese Government. In the symposium, key issues discussed by participants included: transportation of water into the deep mantle and its role in slab-related dynamics; observational and experimental constraints on deep slab properties and the slab environment; modeling of slab stagnation to constrain its mechanisms in comparison with observational and experimental data; observational, experimental and modeling constraints on the fate of stagnant slabs; eventual accumulation of stagnant slabs on the core-mantle boundary and its geodynamic implications. This special issue is a collection of papers presented in the symposium and other papers related to the subject of the symposium. The collected papers provide an overview of the wide range of multidisciplinary studies of mantle dynamics, particularly in the context of subduction, stagnation, and the fate of deep slabs.

  16. Training Deep Spiking Neural Networks Using Backpropagation.

    Science.gov (United States)

    Lee, Jun Haeng; Delbruck, Tobi; Pfeiffer, Michael

    2016-01-01

    Deep spiking neural networks (SNNs) hold the potential for improving the latency and energy efficiency of deep neural networks through data-driven event-based computation. However, training such networks is difficult due to the non-differentiable nature of spike events. In this paper, we introduce a novel technique, which treats the membrane potentials of spiking neurons as differentiable signals, where discontinuities at spike times are considered as noise. This enables an error backpropagation mechanism for deep SNNs that follows the same principles as in conventional deep networks, but works directly on spike signals and membrane potentials. Compared with previous methods relying on indirect training and conversion, our technique has the potential to capture the statistics of spikes more precisely. We evaluate the proposed framework on artificially generated events from the original MNIST handwritten digit benchmark, and also on the N-MNIST benchmark recorded with an event-based dynamic vision sensor, in which the proposed method reduces the error rate by a factor of more than three compared to the best previous SNN, and also achieves a higher accuracy than a conventional convolutional neural network (CNN) trained and tested on the same data. We demonstrate in the context of the MNIST task that thanks to their event-driven operation, deep SNNs (both fully connected and convolutional) trained with our method achieve accuracy equivalent with conventional neural networks. In the N-MNIST example, equivalent accuracy is achieved with about five times fewer computational operations.

  17. Deep Ocean Contribution to Sea Level Rise

    Science.gov (United States)

    Chang, L.; Sun, W.; Tang, H.; Wang, Q.

    2017-12-01

    The ocean temperature and salinity change in the upper 2000m can be detected by Argo floats, so we can know the steric height change of the ocean. But the ocean layers above 2000m represent only 50% of the total ocean volume. Although the temperature and salinity change are small compared to the upper ocean, the deep ocean contribution to sea level might be significant because of its large volume. There has been some research on the deep ocean rely on the very sparse situ observation and are limited to decadal and longer-term rates of change. The available observational data in the deep ocean are too spares to determine the temporal variability, and the long-term changes may have a bias. We will use the Argo date and combine the situ data and topographic data to estimate the temperature and salinity of the sea water below 2000m, so we can obtain a monthly data. We will analyze the seasonal and annual change of the steric height change due to the deep ocean between 2005 and 2016. And we will evaluate the result combination the present-day satellite and in situ observing systems. The deep ocean contribution can be inferred indirectly as the difference between the altimetry minus GRACE and Argo-based steric sea level.

  18. Deep Learning: A Primer for Radiologists.

    Science.gov (United States)

    Chartrand, Gabriel; Cheng, Phillip M; Vorontsov, Eugene; Drozdzal, Michal; Turcotte, Simon; Pal, Christopher J; Kadoury, Samuel; Tang, An

    2017-01-01

    Deep learning is a class of machine learning methods that are gaining success and attracting interest in many domains, including computer vision, speech recognition, natural language processing, and playing games. Deep learning methods produce a mapping from raw inputs to desired outputs (eg, image classes). Unlike traditional machine learning methods, which require hand-engineered feature extraction from inputs, deep learning methods learn these features directly from data. With the advent of large datasets and increased computing power, these methods can produce models with exceptional performance. These models are multilayer artificial neural networks, loosely inspired by biologic neural systems. Weighted connections between nodes (neurons) in the network are iteratively adjusted based on example pairs of inputs and target outputs by back-propagating a corrective error signal through the network. For computer vision tasks, convolutional neural networks (CNNs) have proven to be effective. Recently, several clinical applications of CNNs have been proposed and studied in radiology for classification, detection, and segmentation tasks. This article reviews the key concepts of deep learning for clinical radiologists, discusses technical requirements, describes emerging applications in clinical radiology, and outlines limitations and future directions in this field. Radiologists should become familiar with the principles and potential applications of deep learning in medical imaging. © RSNA, 2017.

  19. DeepPVP: phenotype-based prioritization of causative variants using deep learning

    KAUST Repository

    Boudellioua, Imene

    2018-05-02

    Background: Prioritization of variants in personal genomic data is a major challenge. Recently, computational methods that rely on comparing phenotype similarity have shown to be useful to identify causative variants. In these methods, pathogenicity prediction is combined with a semantic similarity measure to prioritize not only variants that are likely to be dysfunctional but those that are likely involved in the pathogenesis of a patient\\'s phenotype. Results: We have developed DeepPVP, a variant prioritization method that combined automated inference with deep neural networks to identify the likely causative variants in whole exome or whole genome sequence data. We demonstrate that DeepPVP performs significantly better than existing methods, including phenotype-based methods that use similar features. DeepPVP is freely available at https://github.com/bio-ontology-research-group/phenomenet-vp Conclusions: DeepPVP further improves on existing variant prioritization methods both in terms of speed as well as accuracy.

  20. Deep learning in TMVA Benchmarking Benchmarking TMVA DNN Integration of a Deep Autoencoder

    CERN Document Server

    Huwiler, Marc

    2017-01-01

    The TMVA library in ROOT is dedicated to multivariate analysis, and in partic- ular oers numerous machine learning algorithms in a standardized framework. It is widely used in High Energy Physics for data analysis, mainly to perform regression and classication. To keep up to date with the state of the art in deep learning, a new deep learning module was being developed this summer, oering deep neural net- work, convolutional neural network, and autoencoder. TMVA did not have yet any autoencoder method, and the present project consists in implementing the TMVA autoencoder class based on the deep learning module. It also includes some bench- marking performed on the actual deep neural network implementation, in comparison to the Keras framework with Tensorflow and Theano backend.

  1. Deep Space Network Antenna Logic Controller

    Science.gov (United States)

    Ahlstrom, Harlow; Morgan, Scott; Hames, Peter; Strain, Martha; Owen, Christopher; Shimizu, Kenneth; Wilson, Karen; Shaller, David; Doktomomtaz, Said; Leung, Patrick

    2007-01-01

    The Antenna Logic Controller (ALC) software controls and monitors the motion control equipment of the 4,000-metric-ton structure of the Deep Space Network 70-meter antenna. This program coordinates the control of 42 hydraulic pumps, while monitoring several interlocks for personnel and equipment safety. Remote operation of the ALC runs via the Antenna Monitor & Control (AMC) computer, which orchestrates the tracking functions of the entire antenna. This software provides a graphical user interface for local control, monitoring, and identification of faults as well as, at a high level, providing for the digital control of the axis brakes so that the servo of the AMC may control the motion of the antenna. Specific functions of the ALC also include routines for startup in cold weather, controlled shutdown for both normal and fault situations, and pump switching on failure. The increased monitoring, the ability to trend key performance characteristics, the improved fault detection and recovery, the centralization of all control at a single panel, and the simplification of the user interface have all reduced the required workforce to run 70-meter antennas. The ALC also increases the antenna availability by reducing the time required to start up the antenna, to diagnose faults, and by providing additional insight into the performance of key parameters that aid in preventive maintenance to avoid key element failure. The ALC User Display (AUD) is a graphical user interface with hierarchical display structure, which provides high-level status information to the operation of the ALC, as well as detailed information for virtually all aspects of the ALC via drill-down displays. The operational status of an item, be it a function or assembly, is shown in the higher-level display. By pressing the item on the display screen, a new screen opens to show more detail of the function/assembly. Navigation tools and the map button allow immediate access to all screens.

  2. Deep inelastic collisions between very heavy nuclei

    International Nuclear Information System (INIS)

    Sann, H.; Olmi, A.; Civelekoglu, Y.

    1977-01-01

    A systematic survey of deep inelastic reactions was performed for colliding nuclei of masses between 80 and 240 amu. The application of large surface detectors and, particularly, of a position sensitive ionization chamber, has proved to be very effective and appropriate for this type of investigation. The Wilczynski diagrams describing the relative motion between the colliding objects shows a gradual trend as a function of growing masses of target and projectile where the trajectories lead the particles not toward negative scattering angles but increasingly into the direction around and above the grazing angle. This behavior is attributed to a delicate balance between Coulomb and nuclear forces. The energy dumping as a function of the mass transfer strength matches a general law between total kinetic energy loss and the variance of the proton number distribution. For the partly damped component this relation seems to hold independently from the choice of ingoing channel and bombarding energy. The dissipation of the kinetic energy does not depend only on the relative velocity of the impinging nuclei, and the simple friction model is not appropriate to describe these processes. The γ-multiplicity measurement displays a rapid increase as a function of scattering angle and total kinetic energy loss, which give new insights to the process and indicate the necessity of microscopic quantum mechanical calculations of the interaction. In the U-U collision large mass transfers are present which possibly populate with relatively large cross sections the transuranic elements. In the Pb-Pb reaction the mass transfer is more restricted. The decay probability by fission of the primary masses increases strongly for growing masses and excitation energies

  3. Precipitation Nowcast using Deep Recurrent Neural Network

    Science.gov (United States)

    Akbari Asanjan, A.; Yang, T.; Gao, X.; Hsu, K. L.; Sorooshian, S.

    2016-12-01

    An accurate precipitation nowcast (0-6 hours) with a fine temporal and spatial resolution has always been an important prerequisite for flood warning, streamflow prediction and risk management. Most of the popular approaches used for forecasting precipitation can be categorized into two groups. One type of precipitation forecast relies on numerical modeling of the physical dynamics of atmosphere and another is based on empirical and statistical regression models derived by local hydrologists or meteorologists. Given the recent advances in artificial intelligence, in this study a powerful Deep Recurrent Neural Network, termed as Long Short-Term Memory (LSTM) model, is creatively used to extract the patterns and forecast the spatial and temporal variability of Cloud Top Brightness Temperature (CTBT) observed from GOES satellite. Then, a 0-6 hours precipitation nowcast is produced using a Precipitation Estimation from Remote Sensing Information using Artificial Neural Network (PERSIANN) algorithm, in which the CTBT nowcast is used as the PERSIANN algorithm's raw inputs. Two case studies over the continental U.S. have been conducted that demonstrate the improvement of proposed approach as compared to a classical Feed Forward Neural Network and a couple simple regression models. The advantages and disadvantages of the proposed method are summarized with regard to its capability of pattern recognition through time, handling of vanishing gradient during model learning, and working with sparse data. The studies show that the LSTM model performs better than other methods, and it is able to learn the temporal evolution of the precipitation events through over 1000 time lags. The uniqueness of PERSIANN's algorithm enables an alternative precipitation nowcast approach as demonstrated in this study, in which the CTBT prediction is produced and used as the inputs for generating precipitation nowcast.

  4. Deep convective clouds at the tropopause

    Directory of Open Access Journals (Sweden)

    H. H. Aumann

    2011-02-01

    Full Text Available Data from the Atmospheric Infrared Sounder (AIRS on the EOS Aqua spacecraft each day show tens of thousands of Cold Clouds (CC in the tropical oceans with 10 μm window channel brightness temperatures colder than 225 K. These clouds represent a mix of cold anvil clouds and Deep Convective Clouds (DCC. This mix can be separated by computing the difference between two channels, a window channel and a channel with strong CO2 absorption: for some cold clouds this difference is negative, i.e. the spectra for some cold clouds are inverted. We refer to cold clouds with spectra which are more than 2 K inverted as DCCi2. Associated with DCCi2 is a very high rain rate and a local upward displacement of the tropopause, a cold "bulge", which can be seen directly in the brightness temperatures of AIRS and Advanced Microwave Sounding Unit (AMSU temperature sounding channels in the lower stratosphere. The very high rain rate and the local distortion of the tropopause indicate that DCCi2 objects are associated with severe storms. Significant long-term trends in the statistical properties of DCCi2 could be interesting indicators of climate change. While the analysis of the nature and physical conditions related to DCCi2 requires hyperspectral infrared and microwave data, the identification of DCCi2 requires only one good window channel and one strong CO2 sounding channel. This suggests that improved identification of severe storms with future advanced geostationary satellites could be accomplished with the addition of one or two narrow band channels.

  5. Large-eddy simulation of maritime deep tropical convection

    Directory of Open Access Journals (Sweden)

    Peter A Bogenschutz

    2009-12-01

    Full Text Available This study represents an attempt to apply Large-Eddy Simulation (LES resolution to simulate deep tropical convection in near equilibrium for 24 hours over an area of about 205 x 205 km2, which is comparable to that of a typical horizontal grid cell in a global climate model. The simulation is driven by large-scale thermodynamic tendencies derived from mean conditions during the GATE Phase III field experiment. The LES uses 2048 x 2048 x 256 grid points with horizontal grid spacing of 100 m and vertical grid spacing ranging from 50 m in the boundary layer to 100 m in the free troposphere. The simulation reaches a near equilibrium deep convection regime in 12 hours. The simulated vertical cloud distribution exhibits a trimodal vertical distribution of deep, middle and shallow clouds similar to that often observed in Tropics. A sensitivity experiment in which cold pools are suppressed by switching off the evaporation of precipitation results in much lower amounts of shallow and congestus clouds. Unlike the benchmark LES where the new deep clouds tend to appear along the edges of spreading cold pools, the deep clouds in the no-cold-pool experiment tend to reappear at the sites of the previous deep clouds and tend to be surrounded by extensive areas of sporadic shallow clouds. The vertical velocity statistics of updraft and downdraft cores below 6 km height are compared to aircraft observations made during GATE. The comparison shows generally good agreement, and strongly suggests that the LES simulation can be used as a benchmark to represent the dynamics of tropical deep convection on scales ranging from large turbulent eddies to mesoscale convective systems. The effect of horizontal grid resolution is examined by running the same case with progressively larger grid sizes of 200, 400, 800, and 1600 m. These runs show a reasonable agreement with the benchmark LES in statistics such as convective available potential energy, convective inhibition

  6. Deep Learning in Open Source Learning Streams

    DEFF Research Database (Denmark)

    Kjærgaard, Thomas

    2016-01-01

    This chapter presents research on deep learning in a digital learning environment and raises the question if digital instructional designs can catalyze deeper learning than traditional classroom teaching. As a theoretical point of departure the notion of ‘situated learning’ is utilized...... and contrasted to the notion of functionalistic learning in a digital context. The mechanism that enables deep learning in this context is ‘The Open Source Learning Stream’. ‘The Open Source Learning Stream’ is the notion of sharing ‘learning instances’ in a digital space (discussion board, Facebook group......, unistructural, multistructural or relational learning. The research concludes that ‘The Open Source Learning Stream’ can catalyze deep learning and that there are four types of ‘Open Source Learning streams’; individual/ asynchronous, individual/synchronous, shared/asynchronous and shared...

  7. Deep learning in medical imaging: General overview

    Energy Technology Data Exchange (ETDEWEB)

    Lee, June Goo; Jun, Sang Hoon; Cho, Young Won; Lee, Hyun Na; KIm, Guk Bae; Seo, Joon Beom; Kim, Nam Kug [University of Ulsan College of Medicine, Asan Medical Center, Seoul (Korea, Republic of)

    2017-08-01

    The artificial neural network (ANN)–a machine learning technique inspired by the human neuronal synapse system–was introduced in the 1950s. However, the ANN was previously limited in its ability to solve actual problems, due to the vanishing gradient and overfitting problems with training of deep architecture, lack of computing power, and primarily the absence of sufficient data to train the computer system. Interest in this concept has lately resurfaced, due to the availability of big data, enhanced computing power with the current graphics processing units, and novel algorithms to train the deep neural network. Recent studies on this technology suggest its potentially to perform better than humans in some visual and auditory recognition tasks, which may portend its applications in medicine and health care, especially in medical imaging, in the foreseeable future. This review article offers perspectives on the history, development, and applications of deep learning technology, particularly regarding its applications in medical imaging.

  8. Strategic Technologies for Deep Space Transport

    Science.gov (United States)

    Litchford, Ronald J.

    2016-01-01

    Deep space transportation capability for science and exploration is fundamentally limited by available propulsion technologies. Traditional chemical systems are performance plateaued and require enormous Initial Mass in Low Earth Orbit (IMLEO) whereas solar electric propulsion systems are power limited and unable to execute rapid transits. Nuclear based propulsion and alternative energetic methods, on the other hand, represent potential avenues, perhaps the only viable avenues, to high specific power space transport evincing reduced trip time, reduced IMLEO, and expanded deep space reach. Here, key deep space transport mission capability objectives are reviewed in relation to STMD technology portfolio needs, and the advanced propulsion technology solution landscape is examined including open questions, technical challenges, and developmental prospects. Options for potential future investment across the full compliment of STMD programs are presented based on an informed awareness of complimentary activities in industry, academia, OGAs, and NASA mission directorates.

  9. Deep learning in medical imaging: General overview

    International Nuclear Information System (INIS)

    Lee, June Goo; Jun, Sang Hoon; Cho, Young Won; Lee, Hyun Na; KIm, Guk Bae; Seo, Joon Beom; Kim, Nam Kug

    2017-01-01

    The artificial neural network (ANN)–a machine learning technique inspired by the human neuronal synapse system–was introduced in the 1950s. However, the ANN was previously limited in its ability to solve actual problems, due to the vanishing gradient and overfitting problems with training of deep architecture, lack of computing power, and primarily the absence of sufficient data to train the computer system. Interest in this concept has lately resurfaced, due to the availability of big data, enhanced computing power with the current graphics processing units, and novel algorithms to train the deep neural network. Recent studies on this technology suggest its potentially to perform better than humans in some visual and auditory recognition tasks, which may portend its applications in medicine and health care, especially in medical imaging, in the foreseeable future. This review article offers perspectives on the history, development, and applications of deep learning technology, particularly regarding its applications in medical imaging

  10. Oceanography related to deep sea waste disposal

    International Nuclear Information System (INIS)

    1978-09-01

    In connection with studies on the feasibility of the safe disposal of radioactive waste, from a large scale nuclear power programme, either on the bed of the deep ocean or within the deep ocean bed, preparation of the present document was commissioned by the (United Kingdom) Department of the Environment. It attempts (a) to summarize the present state of knowledge of the deep ocean environment relevant to the disposal options and assess the processes which could aid or hinder dispersal of material released from its container; (b) to identify areas of research in which more work is needed before the safety of disposal on, or beneath, the ocean bed can be assessed; and (c) to indicate which areas of research can or should be undertaken by British scientists. The programmes of international cooperation in this field are discussed. The report is divided into four chapters dealing respectively with geology and geophysics, geochemistry, physical oceanography and marine biology. (U.K.)

  11. In Brief: Deep-sea observatory

    Science.gov (United States)

    Showstack, Randy

    2008-11-01

    The first deep-sea ocean observatory offshore of the continental United States has begun operating in the waters off central California. The remotely operated Monterey Accelerated Research System (MARS) will allow scientists to monitor the deep sea continuously. Among the first devices to be hooked up to the observatory are instruments to monitor earthquakes, videotape deep-sea animals, and study the effects of acidification on seafloor animals. ``Some day we may look back at the first packets of data streaming in from the MARS observatory as the equivalent of those first words spoken by Alexander Graham Bell: `Watson, come here, I need you!','' commented Marcia McNutt, president and CEO of the Monterey Bay Aquarium Research Institute, which coordinated construction of the observatory. For more information, see http://www.mbari.org/news/news_releases/2008/mars-live/mars-live.html.

  12. Deep Learning in Medical Imaging: General Overview

    Science.gov (United States)

    Lee, June-Goo; Jun, Sanghoon; Cho, Young-Won; Lee, Hyunna; Kim, Guk Bae

    2017-01-01

    The artificial neural network (ANN)–a machine learning technique inspired by the human neuronal synapse system–was introduced in the 1950s. However, the ANN was previously limited in its ability to solve actual problems, due to the vanishing gradient and overfitting problems with training of deep architecture, lack of computing power, and primarily the absence of sufficient data to train the computer system. Interest in this concept has lately resurfaced, due to the availability of big data, enhanced computing power with the current graphics processing units, and novel algorithms to train the deep neural network. Recent studies on this technology suggest its potentially to perform better than humans in some visual and auditory recognition tasks, which may portend its applications in medicine and healthcare, especially in medical imaging, in the foreseeable future. This review article offers perspectives on the history, development, and applications of deep learning technology, particularly regarding its applications in medical imaging. PMID:28670152

  13. Deep Learning in Medical Image Analysis.

    Science.gov (United States)

    Shen, Dinggang; Wu, Guorong; Suk, Heung-Il

    2017-06-21

    This review covers computer-assisted analysis of images in the field of medical imaging. Recent advances in machine learning, especially with regard to deep learning, are helping to identify, classify, and quantify patterns in medical images. At the core of these advances is the ability to exploit hierarchical feature representations learned solely from data, instead of features designed by hand according to domain-specific knowledge. Deep learning is rapidly becoming the state of the art, leading to enhanced performance in various medical applications. We introduce the fundamentals of deep learning methods and review their successes in image registration, detection of anatomical and cellular structures, tissue segmentation, computer-aided disease diagnosis and prognosis, and so on. We conclude by discussing research issues and suggesting future directions for further improvement.

  14. Pathways to deep decarbonization - Interim 2014 Report

    International Nuclear Information System (INIS)

    2014-01-01

    The interim 2014 report by the Deep Decarbonization Pathways Project (DDPP), coordinated and published by IDDRI and the Sustainable Development Solutions Network (SDSN), presents preliminary findings of the pathways developed by the DDPP Country Research Teams with the objective of achieving emission reductions consistent with limiting global warming to less than 2 deg. C. The DDPP is a knowledge network comprising 15 Country Research Teams and several Partner Organizations who develop and share methods, assumptions, and findings related to deep decarbonization. Each DDPP Country Research Team has developed an illustrative road-map for the transition to a low-carbon economy, with the intent of taking into account national socio-economic conditions, development aspirations, infrastructure stocks, resource endowments, and other relevant factors. The interim 2014 report focuses on technically feasible pathways to deep decarbonization

  15. Excess plutonium disposition: The deep borehole option

    International Nuclear Information System (INIS)

    Ferguson, K.L.

    1994-01-01

    This report reviews the current status of technologies required for the disposition of plutonium in Very Deep Holes (VDH). It is in response to a recent National Academy of Sciences (NAS) report which addressed the management of excess weapons plutonium and recommended three approaches to the ultimate disposition of excess plutonium: (1) fabrication and use as a fuel in existing or modified reactors in a once-through cycle, (2) vitrification with high-level radioactive waste for repository disposition, (3) burial in deep boreholes. As indicated in the NAS report, substantial effort would be required to address the broad range of issues related to deep bore-hole emplacement. Subjects reviewed in this report include geology and hydrology, design and engineering, safety and licensing, policy decisions that can impact the viability of the concept, and applicable international programs. Key technical areas that would require attention should decisions be made to further develop the borehole emplacement option are identified

  16. Deep Learning in Medical Imaging: General Overview.

    Science.gov (United States)

    Lee, June-Goo; Jun, Sanghoon; Cho, Young-Won; Lee, Hyunna; Kim, Guk Bae; Seo, Joon Beom; Kim, Namkug

    2017-01-01

    The artificial neural network (ANN)-a machine learning technique inspired by the human neuronal synapse system-was introduced in the 1950s. However, the ANN was previously limited in its ability to solve actual problems, due to the vanishing gradient and overfitting problems with training of deep architecture, lack of computing power, and primarily the absence of sufficient data to train the computer system. Interest in this concept has lately resurfaced, due to the availability of big data, enhanced computing power with the current graphics processing units, and novel algorithms to train the deep neural network. Recent studies on this technology suggest its potentially to perform better than humans in some visual and auditory recognition tasks, which may portend its applications in medicine and healthcare, especially in medical imaging, in the foreseeable future. This review article offers perspectives on the history, development, and applications of deep learning technology, particularly regarding its applications in medical imaging.

  17. L-proline-based deep eutectic solvents (DESs) for deep catalytic oxidative desulfurization (ODS) of diesel.

    Science.gov (United States)

    Hao, Lingwan; Wang, Meiri; Shan, Wenjuan; Deng, Changliang; Ren, Wanzhong; Shi, Zhouzhou; Lü, Hongying

    2017-10-05

    A series of L-proline-based DESs was prepared through an atom economic reaction between L-proline (L-Pro) and four different kinds of organic acids. The DESs were characterized by Fourier transform infrared spectroscopy (FT-IR), H nuclear magnetic resonance ( 1 HNMR), cyclic voltammogram (CV) and the Hammett method. The synthesized DESs were used for the oxidative desulfurization and the L-Pro/p-toluenesultonic acid (L-Pro/p-TsOH) system shows the highest catalytic activity that the removal of dibenzothiophene (DBT) reached 99% at 60°C in 2h, which may involve the dual activation of the L-Pro/p-TsOH. The acidity of four different L-proline-based DESs was measured and the results show that it could not simply conclude that the correlation between the acidity of DESs and desulfurization capability was positive or negative. The electrochemical measurements evidences and recycling experiment indicate a good stability performance of L-Pro/p-TsOH in desulfurization. This work will provide a novel and potential method for the deep oxidation desulfurization. Copyright © 2017 Elsevier B.V. All rights reserved.

  18. Kernel methods for deep learning

    OpenAIRE

    Cho, Youngmin

    2012-01-01

    We introduce a new family of positive-definite kernels that mimic the computation in large neural networks. We derive the different members of this family by considering neural networks with different activation functions. Using these kernels as building blocks, we also show how to construct other positive-definite kernels by operations such as composition, multiplication, and averaging. We explore the use of these kernels in standard models of supervised learning, such as support vector mach...

  19. Deep Space Detection of Oriented Ice Crystals

    Science.gov (United States)

    Marshak, A.; Varnai, T.; Kostinski, A. B.

    2017-12-01

    The deep space climate observatory (DSCOVR) spacecraft resides at the first Lagrangian point about one million miles from Earth. A polychromatic imaging camera onboard delivers nearly hourly observations of the entire sun-lit face of the Earth. Many images contain unexpected bright flashes of light over both ocean and land. We constructed a yearlong time series of flash latitudes, scattering angles and oxygen absorption to demonstrate conclusively that the flashes over land are specular reflections off tiny ice crystals floating in the air nearly horizontally. Such deep space detection of tropospheric ice can be used to constrain the likelihood of oriented crystals and their contribution to Earth albedo.

  20. A clinical study on deep neck abscess

    International Nuclear Information System (INIS)

    Ota, Yumi; Ogawa, Yoshiko; Takemura, Teiji; Sawada, Toru

    2007-01-01

    Although various effective antibiotics have been synthesized, deep neck abscess is still a serious and life-threatening infection. It is important to diagnose promptly and treat adequately, and contrast-enhanced CT is useful and indispensable for diagnosis. We reviewed our patients with deep neck abscess, and analyzed the location by reviewing CT images, and discussed the treatment. Surgical drainage is a fundamental treatment for abscess but if it exists in only one area such as the parotid gland space, it can be cured with needle aspiration and suitable antibiotics. (author)