Challenges in Food Scientist Training in a global setting

Directory of Open Access Journals (Sweden)

Andreas Höhl

2012-10-01

Full Text Available Normal 0 21 false false false EN-GB X-NONE X-NONE Education and training were an integral part of the MoniQA Network of Excellence. Embedded in the "Spreading of excellence programme", Work Package 9 (Joint education programmes and training tools was responsible for establishing a joint training programme for food safety and quality within and beyond the network. So-called `MoniQA Food Scientist Training' (MoniQA FST was offered to provide technical knowledge on different levels and research management skills as well. Training needs for different regions as well as for different target groups (scientists, industry personnel, authorities had to be considered as well as developing strong collaboration links between network partners and related projects. Beside face-to-face workshops e-learning modules have been developed and web seminars were organized. In order to achieve high quality training, a quality assurance concept has been implemented. It turned out that these types of training are of high value in terms of bringing together scientists from different regions and cultures of the globe, involving highly qualified trainers as basis for a sustainable network in the future.

Adaptive training of feedforward neural networks by Kalman filtering

International Nuclear Information System (INIS)

Ciftcioglu, Oe.

1995-02-01

Adaptive training of feedforward neural networks by Kalman filtering is described. Adaptive training is particularly important in estimation by neural network in real-time environmental where the trained network is used for system estimation while the network is further trained by means of the information provided by the experienced/exercised ongoing operation. As result of this, neural network adapts itself to a changing environment to perform its mission without recourse to re-training. The performance of the training method is demonstrated by means of actual process signals from a nuclear power plant. (orig.)

An Improved Neural Network Training Algorithm for Wi-Fi Fingerprinting Positioning

Directory of Open Access Journals (Sweden)

Esmond Mok

2013-09-01

Full Text Available Ubiquitous positioning provides continuous positional information in both indoor and outdoor environments for a wide spectrum of location based service (LBS applications. With the rapid development of the low-cost and high speed data communication, Wi-Fi networks in many metropolitan cities, strength of signals propagated from the Wi-Fi access points (APs namely received signal strength (RSS have been cleverly adopted for indoor positioning. In this paper, a Wi-Fi positioning algorithm based on neural network modeling of Wi-Fi signal patterns is proposed. This algorithm is based on the correlation between the initial parameter setting for neural network training and output of the mean square error to obtain better modeling of the nonlinear highly complex Wi-Fi signal power propagation surface. The test results show that this neural network based data processing algorithm can significantly improve the neural network training surface to achieve the highest possible accuracy of the Wi-Fi fingerprinting positioning method.

Introduction to the EC's Marie Curie Initial Training Network (MC-ITN) project: Particle Training Network for European Radiotherapy (PARTNER).

Science.gov (United States)

Dosanjh, Manjit; Magrin, Giulio

2013-07-01

PARTNER (Particle Training Network for European Radiotherapy) is a project funded by the European Commission's Marie Curie-ITN funding scheme through the ENLIGHT Platform for 5.6 million Euro. PARTNER has brought together academic institutes, research centres and leading European companies, focusing in particular on a specialized radiotherapy (RT) called hadron therapy (HT), interchangeably referred to as particle therapy (PT). The ultimate goal of HT is to deliver more effective treatment to cancer patients leading to major improvement in the health of citizens. In Europe, several hundred million Euro have been invested, since the beginning of this century, in PT. In this decade, the use of HT is rapidly growing across Europe, and there is an urgent need for qualified researchers from a range of disciplines to work on its translational research. In response to this need, the European community of HT, and in particular 10 leading academic institutes, research centres, companies and small and medium-sized enterprises, joined together to form the PARTNER consortium. All partners have international reputations in the diverse but complementary fields associated with PT: clinical, radiobiological and technological. Thus the network incorporates a unique set of competencies, expertise, infrastructures and training possibilities. This paper describes the status and needs of PT research in Europe, the importance of and challenges associated with the creation of a training network, the objectives, the initial results, and the expected long-term benefits of the PARTNER initiative.

Introduction to the EC's marie curie initial training network (MC-ITN) project. Particle training network for European radiotherapy (PARTNER)

International Nuclear Information System (INIS)

Dosanjh, Manjit; Magrin, Giulio

2013-01-01

PARTNER (Particle Training Network for European Radiotherapy) is a project funded by the European Commission's Marie Curie-ITN funding scheme through the ENLIGHT Platform for 5.6 million Euro. PARTNER has brought together academic institutes, research centres and leading European companies, focusing in particular on a specialized radiotherapy (RT) called hadron therapy (HT), interchangeably referred to as particle therapy (PT). The ultimate goal of HT is to deliver more effective treatment to cancer patients leading to major improvement in the health of citizens. In Europe, several hundred million Euro have been invested, since the beginning of this century, in PT. In this decade, the use of HT is rapidly growing across Europe, and there is an urgent need for qualified researchers from a range of disciplines to work on its translational research. In response to this need, the European community of HT, and in particular 10 leading academic institutes, research centres, companies and small and medium-sized enterprises, joined together to form the PARTNER consortium. All partners have international reputations in the diverse but complementary fields associated with PT: clinical, radiobiological and technological. Thus the network incorporates a unique set of competencies, expertise, infrastructures and training possibilities. This paper describes the status and needs of PT research in Europe, the importance of and challenges associated with the creation of a training network, the objectives, the initial results, and the expected long-term benefits of the PARTNER initiative. (author)

Revisiting Social Network Utilization by Physicians-in-Training.

Science.gov (United States)

Black, Erik W; Thompson, Lindsay A; Duff, W Patrick; Dawson, Kara; Saliba, Heidi; Black, Nicole M Paradise

2010-06-01

To measure and compare the frequency and content of online social networking among 2 cohorts of medical students and residents (2007 and 2009). Using the online social networking application Facebook, we evaluated social networking profiles for 2 cohorts of medical students (n  =  528) and residents (n  =  712) at the University of Florida in Gainesville. Objective measures included existence of a profile, whether it was made private, and whether any personally identifiable information was included. Subjective outcomes included photographic content, affiliated social groups, and personal information not generally disclosed in a doctor-patient encounter. We compared our results to our previously published and reported data from 2007. Social networking continues to be common amongst physicians-in-training, with 39.8% of residents and 69.5% of medical students maintaining Facebook accounts. Residents' participation significantly increased (P privacy settings (P privacy and the expansive and impersonal networks of online "friends" who may view profiles.

Convolutional Neural Networks for Medical Image Analysis: Full Training or Fine Tuning?

OpenAIRE

Tajbakhsh, Nima; Shin, Jae Y.; Gurudu, Suryakanth R.; Hurst, R. Todd; Kendall, Christopher B.; Gotway, Michael B.; Liang, Jianming

2017-01-01

Training a deep convolutional neural network (CNN) from scratch is difficult because it requires a large amount of labeled training data and a great deal of expertise to ensure proper convergence. A promising alternative is to fine-tune a CNN that has been pre-trained using, for instance, a large set of labeled natural images. However, the substantial differences between natural and medical images may advise against such knowledge transfer. In this paper, we seek to answer the following centr...

Transfer Learning for Video Recognition with Scarce Training Data for Deep Convolutional Neural Network

OpenAIRE

Su, Yu-Chuan; Chiu, Tzu-Hsuan; Yeh, Chun-Yen; Huang, Hsin-Fu; Hsu, Winston H.

2014-01-01

Unconstrained video recognition and Deep Convolution Network (DCN) are two active topics in computer vision recently. In this work, we apply DCNs as frame-based recognizers for video recognition. Our preliminary studies, however, show that video corpora with complete ground truth are usually not large and diverse enough to learn a robust model. The networks trained directly on the video data set suffer from significant overfitting and have poor recognition rate on the test set. The same lack-...

Data Programming: Creating Large Training Sets, Quickly

Science.gov (United States)

Ratner, Alexander; De Sa, Christopher; Wu, Sen; Selsam, Daniel; Ré, Christopher

2018-01-01

Large labeled training sets are the critical building blocks of supervised learning methods and are key enablers of deep learning techniques. For some applications, creating labeled training sets is the most time-consuming and expensive part of applying machine learning. We therefore propose a paradigm for the programmatic creation of training sets called data programming in which users express weak supervision strategies or domain heuristics as labeling functions, which are programs that label subsets of the data, but that are noisy and may conflict. We show that by explicitly representing this training set labeling process as a generative model, we can “denoise” the generated training set, and establish theoretically that we can recover the parameters of these generative models in a handful of settings. We then show how to modify a discriminative loss function to make it noise-aware, and demonstrate our method over a range of discriminative models including logistic regression and LSTMs. Experimentally, on the 2014 TAC-KBP Slot Filling challenge, we show that data programming would have led to a new winning score, and also show that applying data programming to an LSTM model leads to a TAC-KBP score almost 6 F1 points over a state-of-the-art LSTM baseline (and into second place in the competition). Additionally, in initial user studies we observed that data programming may be an easier way for non-experts to create machine learning models when training data is limited or unavailable. PMID:29872252

Local Dynamics in Trained Recurrent Neural Networks.

Science.gov (United States)

Rivkind, Alexander; Barak, Omri

2017-06-23

Learning a task induces connectivity changes in neural circuits, thereby changing their dynamics. To elucidate task-related neural dynamics, we study trained recurrent neural networks. We develop a mean field theory for reservoir computing networks trained to have multiple fixed point attractors. Our main result is that the dynamics of the network's output in the vicinity of attractors is governed by a low-order linear ordinary differential equation. The stability of the resulting equation can be assessed, predicting training success or failure. As a consequence, networks of rectified linear units and of sigmoidal nonlinearities are shown to have diametrically different properties when it comes to learning attractors. Furthermore, a characteristic time constant, which remains finite at the edge of chaos, offers an explanation of the network's output robustness in the presence of variability of the internal neural dynamics. Finally, the proposed theory predicts state-dependent frequency selectivity in the network response.

Local Dynamics in Trained Recurrent Neural Networks

Science.gov (United States)

Rivkind, Alexander; Barak, Omri

2017-06-01

Learning a task induces connectivity changes in neural circuits, thereby changing their dynamics. To elucidate task-related neural dynamics, we study trained recurrent neural networks. We develop a mean field theory for reservoir computing networks trained to have multiple fixed point attractors. Our main result is that the dynamics of the network's output in the vicinity of attractors is governed by a low-order linear ordinary differential equation. The stability of the resulting equation can be assessed, predicting training success or failure. As a consequence, networks of rectified linear units and of sigmoidal nonlinearities are shown to have diametrically different properties when it comes to learning attractors. Furthermore, a characteristic time constant, which remains finite at the edge of chaos, offers an explanation of the network's output robustness in the presence of variability of the internal neural dynamics. Finally, the proposed theory predicts state-dependent frequency selectivity in the network response.

Artificial Neural Network with Hardware Training and Hardware Refresh

Science.gov (United States)

Duong, Tuan A. (Inventor)

2003-01-01

A neural network circuit is provided having a plurality of circuits capable of charge storage. Also provided is a plurality of circuits each coupled to at least one of the plurality of charge storage circuits and constructed to generate an output in accordance with a neuron transfer function. Each of a plurality of circuits is coupled to one of the plurality of neuron transfer function circuits and constructed to generate a derivative of the output. A weight update circuit updates the charge storage circuits based upon output from the plurality of transfer function circuits and output from the plurality of derivative circuits. In preferred embodiments, separate training and validation networks share the same set of charge storage circuits and may operate concurrently. The validation network has a separate transfer function circuits each being coupled to the charge storage circuits so as to replicate the training network s coupling of the plurality of charge storage to the plurality of transfer function circuits. The plurality of transfer function circuits may be constructed each having a transconductance amplifier providing differential currents combined to provide an output in accordance with a transfer function. The derivative circuits may have a circuit constructed to generate a biased differential currents combined so as to provide the derivative of the transfer function.

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.

Implementation and Outcomes of a Collaborative Multi-Center Network Aimed at Web-Based Cognitive Training - COGWEB Network.

Science.gov (United States)

Tedim Cruz, Vítor; Pais, Joana; Ruano, Luis; Mateus, Cátia; Colunas, Márcio; Alves, Ivânia; Barreto, Rui; Conde, Eduardo; Sousa, Andreia; Araújo, Isabel; Bento, Virgílio; Coutinho, Paula; Rocha, Nelson

2014-01-01

Cognitive care for the most prevalent neurologic and psychiatric conditions will only improve through the implementation of new sustainable approaches. Innovative cognitive training methodologies and collaborative professional networks are necessary evolutions in the mental health sector. The objective of the study was to describe the implementation process and early outcomes of a nationwide multi-organizational network supported on a Web-based cognitive training system (COGWEB). The setting for network implementation was the Portuguese mental health system and the hospital-, academic-, community-based institutions and professionals providing cognitive training. The network started in August 2012, with 16 centers, and was monitored until September 2013 (inclusions were open). After onsite training, all were allowed to use COGWEB in their clinical or research activities. For supervision and maintenance were implemented newsletters, questionnaires, visits and webinars. The following outcomes were prospectively measured: (1) number, (2) type, (3) time to start, and (4) activity state of centers; age, gender, level of education, and medical diagnosis of patients enrolled. The network included 68 professionals from 41 centers, (33/41) 80% clinical, (8/41) 19% nonclinical. A total of 298 patients received cognitive training; 45.3% (n=135) female, mean age 54.4 years (SD 18.7), mean educational level 9.8 years (SD 4.8). The number enrolled each month increased significantly (r=0.6; P=.031). At 12 months, 205 remained on treatment. The major causes of cognitive impairment were: (1) neurodegenerative (115/298, 38.6%), (2) structural brain lesions (63/298, 21.1%), (3) autoimmune (40/298, 13.4%), (4) schizophrenia (30/298, 10.1%), and (5) others (50/298, 16.8%). The comparison of the patient profiles, promoter versus all other clinical centers, showed significant increases in the diversity of causes and spectrums of ages and education. Over its first year, there was a major

Limited Effects of Set Shifting Training in Healthy Older Adults

Directory of Open Access Journals (Sweden)

Petra Grönholm-Nyman

2017-03-01

Full Text Available Our ability to flexibly shift between tasks or task sets declines in older age. As this decline may have adverse effects on everyday life of elderly people, it is of interest to study whether set shifting ability can be trained, and if training effects generalize to other cognitive tasks. Here, we report a randomized controlled trial where healthy older adults trained set shifting with three different set shifting tasks. The training group (n = 17 performed adaptive set shifting training for 5 weeks with three training sessions a week (45 min/session, while the active control group (n = 16 played three different computer games for the same period. Both groups underwent extensive pre- and post-testing and a 1-year follow-up. Compared to the controls, the training group showed significant improvements on the trained tasks. Evidence for near transfer in the training group was very limited, as it was seen only on overall accuracy on an untrained computerized set shifting task. No far transfer to other cognitive functions was observed. One year later, the training group was still better on the trained tasks but the single near transfer effect had vanished. The results suggest that computerized set shifting training in the elderly shows long-lasting effects on the trained tasks but very little benefit in terms of generalization.

Training trajectories by continuous recurrent multilayer networks.

Science.gov (United States)

Leistritz, L; Galicki, M; Witte, H; Kochs, E

2002-01-01

This paper addresses the problem of training trajectories by means of continuous recurrent neural networks whose feedforward parts are multilayer perceptrons. Such networks can approximate a general nonlinear dynamic system with arbitrary accuracy. The learning process is transformed into an optimal control framework where the weights are the controls to be determined. A training algorithm based upon a variational formulation of Pontryagin's maximum principle is proposed for such networks. Computer examples demonstrating the efficiency of the given approach are also presented.

Superimposed Training-Based Channel Estimation for MIMO Relay Networks

Directory of Open Access Journals (Sweden)

Xiaoyan Xu

2012-01-01

Full Text Available We introduce the superimposed training strategy into the multiple-input multiple-output (MIMO amplify-and-forward (AF one-way relay network (OWRN to perform the individual channel estimation at the destination. Through the superposition of a group of additional training vectors at the relay subject to power allocation, the separated estimates of the source-relay and relay-destination channels can be obtained directly at the destination, and the accordance with the two-hop AF strategy can be guaranteed at the same time. The closed-form Bayesian Cramér-Rao lower bound (CRLB is derived for the estimation of two sets of flat-fading MIMO channel under random channel parameters and further exploited to design the optimal training vectors. A specific suboptimal channel estimation algorithm is applied in the MIMO AF OWRN using the optimal training sequences, and the normalized mean square error performance for the estimation is provided to verify the Bayesian CRLB results.

Fault Diagnosis Method of Polymerization Kettle Equipment Based on Rough Sets and BP Neural Network

Directory of Open Access Journals (Sweden)

Shu-zhi Gao

2013-01-01

Full Text Available Polyvinyl chloride (PVC polymerizing production process is a typical complex controlled object, with complexity features, such as nonlinear, multivariable, strong coupling, and large time-delay. Aiming at the real-time fault diagnosis and optimized monitoring requirements of the large-scale key polymerization equipment of PVC production process, a real-time fault diagnosis strategy is proposed based on rough sets theory with the improved discernibility matrix and BP neural networks. The improved discernibility matrix is adopted to reduct the attributes of rough sets in order to decrease the input dimensionality of fault characteristics effectively. Levenberg-Marquardt BP neural network is trained to diagnose the polymerize faults according to the reducted decision table, which realizes the nonlinear mapping from fault symptom set to polymerize fault set. Simulation experiments are carried out combining with the industry history datum to show the effectiveness of the proposed rough set neural networks fault diagnosis method. The proposed strategy greatly increased the accuracy rate and efficiency of the polymerization fault diagnosis system.

Convolutional neural networks based on augmented training samples for synthetic aperture radar target recognition

Science.gov (United States)

Yan, Yue

2018-03-01

A synthetic aperture radar (SAR) automatic target recognition (ATR) method based on the convolutional neural networks (CNN) trained by augmented training samples is proposed. To enhance the robustness of CNN to various extended operating conditions (EOCs), the original training images are used to generate the noisy samples at different signal-to-noise ratios (SNRs), multiresolution representations, and partially occluded images. Then, the generated images together with the original ones are used to train a designed CNN for target recognition. The augmented training samples can contrapuntally improve the robustness of the trained CNN to the covered EOCs, i.e., the noise corruption, resolution variance, and partial occlusion. Moreover, the significantly larger training set effectively enhances the representation capability for other conditions, e.g., the standard operating condition (SOC), as well as the stability of the network. Therefore, better performance can be achieved by the proposed method for SAR ATR. For experimental evaluation, extensive experiments are conducted on the Moving and Stationary Target Acquisition and Recognition dataset under SOC and several typical EOCs.

THE EXPERIENCE OF NETWORKING POSTGRADUATE TRAINING PROGRAMMES

Directory of Open Access Journals (Sweden)

E. A. Teplyashina

2017-01-01

Full Text Available Introduction. Present scientific and innovative education programmes focus on the development of applied research in priority areas of industry, cross-industry and regional development. Implementation of such programs is most effective along with the network organization of the process of training. In accordance with the Federal Law on Education in the Russian Federation, this model of networking as «educational institution – educational organization» is a very convenient form of academic mobility realisation.The aim of the present paper is to analyse the model of interaction of the networking postgraduate training programmes at Krasnoyarsk State Medical University named after Prof. V. F. Voino-Yasenetsky and Medical School of Niigata University (Japan.Methodology and research methods involve theoretical analysis of the scientific outcomes of implementing a networking postgraduate training programme, comparative-teaching method, generalization, and pedagogical modeling.Results. The mechanisms of developing the partnership between universities of different countries are detailed. The experience of network international education in a postgraduate study is presented. The presented experience allowed the authors to develop an integrated strategy of cooperation with foreign colleagues in this direction. The advantages and problems of use of a network form of training of academic and teaching staff in a postgraduate school are revealed. The proposals and recommendations on optimization and harmonization of the purposes, tasks and programs of network interaction of the educational organizations are formulated.Practical significance. The proposed materials of the publication can form the base for creation and designing of an effective system of postgraduate education and competitiveness growth of the Russian universities. 

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.

Global Optimization for Transport Network Expansion and Signal Setting

OpenAIRE

Liu, Haoxiang; Wang, David Z. W.; Yue, Hao

2015-01-01

This paper proposes a model to address an urban transport planning problem involving combined network design and signal setting in a saturated network. Conventional transport planning models usually deal with the network design problem and signal setting problem separately. However, the fact that network capacity design and capacity allocation determined by network signal setting combine to govern the transport network performance requires the optimal transport planning to consider the two pr...

Permitted and forbidden sets in symmetric threshold-linear networks.

Science.gov (United States)

Hahnloser, Richard H R; Seung, H Sebastian; Slotine, Jean-Jacques

2003-03-01

The richness and complexity of recurrent cortical circuits is an inexhaustible source of inspiration for thinking about high-level biological computation. In past theoretical studies, constraints on the synaptic connection patterns of threshold-linear networks were found that guaranteed bounded network dynamics, convergence to attractive fixed points, and multistability, all fundamental aspects of cortical information processing. However, these conditions were only sufficient, and it remained unclear which were the minimal (necessary) conditions for convergence and multistability. We show that symmetric threshold-linear networks converge to a set of attractive fixed points if and only if the network matrix is copositive. Furthermore, the set of attractive fixed points is nonconnected (the network is multiattractive) if and only if the network matrix is not positive semidefinite. There are permitted sets of neurons that can be coactive at a stable steady state and forbidden sets that cannot. Permitted sets are clustered in the sense that subsets of permitted sets are permitted and supersets of forbidden sets are forbidden. By viewing permitted sets as memories stored in the synaptic connections, we provide a formulation of long-term memory that is more general than the traditional perspective of fixed-point attractor networks. There is a close correspondence between threshold-linear networks and networks defined by the generalized Lotka-Volterra equations.

Digital intelligent booster for DCC miniature train networks

Science.gov (United States)

Ursu, M. P.; Condruz, D. A.

2017-08-01

Modern miniature trains are now driven by means of the DCC (Digital Command and Control) system, which allows the human operator or a personal computer to launch commands to each individual train or even to control different features of the same train. The digital command station encodes these commands and sends them to the trains by means of electrical pulses via the rails of the railway network. Due to the development of the miniature railway network, it may happen that the power requirement of the increasing number of digital locomotives, carriages and accessories exceeds the nominal output power of the digital command station. This digital intelligent booster relieves the digital command station from powering the entire railway network all by itself, and it automatically handles the multiple powered sections of the network. This electronic device is also able to detect and process short-circuits and overload conditions, without the intervention of the digital command station.

An Improved Walk Model for Train Movement on Railway Network

International Nuclear Information System (INIS)

Li Keping; Mao Bohua; Gao Ziyou

2009-01-01

In this paper, we propose an improved walk model for simulating the train movement on railway network. In the proposed method, walkers represent trains. The improved walk model is a kind of the network-based simulation analysis model. Using some management rules for walker movement, walker can dynamically determine its departure and arrival times at stations. In order to test the proposed method, we simulate the train movement on a part of railway network. The numerical simulation and analytical results demonstrate that the improved model is an effective tool for simulating the train movement on railway network. Moreover, it can well capture the characteristic behaviors of train scheduling in railway traffic. (general)

Supervised learning in spiking neural networks with FORCE training.

Science.gov (United States)

Nicola, Wilten; Clopath, Claudia

2017-12-20

Populations of neurons display an extraordinary diversity in the behaviors they affect and display. Machine learning techniques have recently emerged that allow us to create networks of model neurons that display behaviors of similar complexity. Here we demonstrate the direct applicability of one such technique, the FORCE method, to spiking neural networks. We train these networks to mimic dynamical systems, classify inputs, and store discrete sequences that correspond to the notes of a song. Finally, we use FORCE training to create two biologically motivated model circuits. One is inspired by the zebra finch and successfully reproduces songbird singing. The second network is motivated by the hippocampus and is trained to store and replay a movie scene. FORCE trained networks reproduce behaviors comparable in complexity to their inspired circuits and yield information not easily obtainable with other techniques, such as behavioral responses to pharmacological manipulations and spike timing statistics.

Settings in Social Networks : a Measurement Model

NARCIS (Netherlands)

Schweinberger, Michael; Snijders, Tom A.B.

2003-01-01

A class of statistical models is proposed that aims to recover latent settings structures in social networks. Settings may be regarded as clusters of vertices. The measurement model is based on two assumptions. (1) The observed network is generated by hierarchically nested latent transitive

MO-DE-BRA-04: The CREATE Medical Physics Research Training Network: Training of New Generation Innovators

Energy Technology Data Exchange (ETDEWEB)

Seuntjens, J; Collins, L; Devic, S; El Naqa, I; Nadeau, J; Reader, A [McGill University, Montreal, QC (Canada); Beaulieu, L; Despres, P [Centre Hospitalier Univ de Quebec, Quebec, QC (Canada); Pike, B [University of Calgary, Calgary, Alberta (Canada)

2015-06-15

Purpose: Over the past century, physicists have played a major role in transforming scientific discovery into everyday clinical applications. However, with the increasingly stringent requirements to regulate medical physics as a health profession, the role of physicists as scientists and innovators has become at serious risk of erosion. These challenges trigger the need for a new, revolutionized training program at the graduate level that respects scientific rigor, attention for medical physics-relevant developments in basic sciences, innovation and entrepreneurship. Methods: A grant proposal was funded by the Collaborative REsearch and Training Experience program (CREATE) of the Natural Sciences and Engineering Research Council (NSERC) of Canada. This enabled the creation of the Medical Physics Research Training Network (MPRTN) around two CAMPEP-accredited medical physics programs. Members of the network consist of medical device companies, government (research and regulatory) and academia. The MPRTN/CREATE program proposes a curriculum with three main themes: (1) radiation physics, (2) imaging & image processing and (3) radiation response, outcomes and modeling. Results: The MPRTN was created mid 2013 (mprtn.com) and features (1) four new basic Ph.D. courses; (2) industry participation in research projects; (3) formal job-readiness training with involvement of guest faculty from academia, government and industry. MPRTN activities since 2013 include 22 conferences; 7 workshops and 4 exchange travels. Three patents were filed or issued, nine awards/best papers were won. Fifteen journal publications were accepted/published, 102 conference abstracts. There are now 13 industry partners. Conclusion: A medical physics research training network has been set up with the goal to harness graduate student’s job-readiness for industry, government and academia in addition to the conventional clinical role. Two years after inception, significant successes have been booked

MO-DE-BRA-04: The CREATE Medical Physics Research Training Network: Training of New Generation Innovators

International Nuclear Information System (INIS)

Seuntjens, J; Collins, L; Devic, S; El Naqa, I; Nadeau, J; Reader, A; Beaulieu, L; Despres, P; Pike, B

2015-01-01

Purpose: Over the past century, physicists have played a major role in transforming scientific discovery into everyday clinical applications. However, with the increasingly stringent requirements to regulate medical physics as a health profession, the role of physicists as scientists and innovators has become at serious risk of erosion. These challenges trigger the need for a new, revolutionized training program at the graduate level that respects scientific rigor, attention for medical physics-relevant developments in basic sciences, innovation and entrepreneurship. Methods: A grant proposal was funded by the Collaborative REsearch and Training Experience program (CREATE) of the Natural Sciences and Engineering Research Council (NSERC) of Canada. This enabled the creation of the Medical Physics Research Training Network (MPRTN) around two CAMPEP-accredited medical physics programs. Members of the network consist of medical device companies, government (research and regulatory) and academia. The MPRTN/CREATE program proposes a curriculum with three main themes: (1) radiation physics, (2) imaging & image processing and (3) radiation response, outcomes and modeling. Results: The MPRTN was created mid 2013 (mprtn.com) and features (1) four new basic Ph.D. courses; (2) industry participation in research projects; (3) formal job-readiness training with involvement of guest faculty from academia, government and industry. MPRTN activities since 2013 include 22 conferences; 7 workshops and 4 exchange travels. Three patents were filed or issued, nine awards/best papers were won. Fifteen journal publications were accepted/published, 102 conference abstracts. There are now 13 industry partners. Conclusion: A medical physics research training network has been set up with the goal to harness graduate student’s job-readiness for industry, government and academia in addition to the conventional clinical role. Two years after inception, significant successes have been booked

Reconstruction of sparse connectivity in neural networks from spike train covariances

International Nuclear Information System (INIS)

Pernice, Volker; Rotter, Stefan

2013-01-01

The inference of causation from correlation is in general highly problematic. Correspondingly, it is difficult to infer the existence of physical synaptic connections between neurons from correlations in their activity. Covariances in neural spike trains and their relation to network structure have been the subject of intense research, both experimentally and theoretically. The influence of recurrent connections on covariances can be characterized directly in linear models, where connectivity in the network is described by a matrix of linear coupling kernels. However, as indirect connections also give rise to covariances, the inverse problem of inferring network structure from covariances can generally not be solved unambiguously. Here we study to what degree this ambiguity can be resolved if the sparseness of neural networks is taken into account. To reconstruct a sparse network, we determine the minimal set of linear couplings consistent with the measured covariances by minimizing the L 1 norm of the coupling matrix under appropriate constraints. Contrary to intuition, after stochastic optimization of the coupling matrix, the resulting estimate of the underlying network is directed, despite the fact that a symmetric matrix of count covariances is used for inference. The performance of the new method is best if connections are neither exceedingly sparse, nor too dense, and it is easily applicable for networks of a few hundred nodes. Full coupling kernels can be obtained from the matrix of full covariance functions. We apply our method to networks of leaky integrate-and-fire neurons in an asynchronous–irregular state, where spike train covariances are well described by a linear model. (paper)

Train Stop Scheduling in a High-Speed Rail Network by Utilizing a Two-Stage Approach

Directory of Open Access Journals (Sweden)

Huiling Fu

2012-01-01

Full Text Available Among the most commonly used methods of scheduling train stops are practical experience and various “one-step” optimal models. These methods face problems of direct transferability and computational complexity when considering a large-scale high-speed rail (HSR network such as the one in China. This paper introduces a two-stage approach for train stop scheduling with a goal of efficiently organizing passenger traffic into a rational train stop pattern combination while retaining features of regularity, connectivity, and rapidity (RCR. Based on a three-level station classification definition, a mixed integer programming model and a train operating tactics descriptive model along with the computing algorithm are developed and presented for the two stages. A real-world numerical example is presented using the Chinese HSR network as the setting. The performance of the train stop schedule and the applicability of the proposed approach are evaluated from the perspective of maintaining RCR.

Neural network training by Kalman filtering in process system monitoring

International Nuclear Information System (INIS)

Ciftcioglu, Oe.

1996-03-01

Kalman filtering approach for neural network training is described. Its extended form is used as an adaptive filter in a nonlinear environment of the form a feedforward neural network. Kalman filtering approach generally provides fast training as well as avoiding excessive learning which results in enhanced generalization capability. The network is used in a process monitoring application where the inputs are measurement signals. Since the measurement errors are also modelled in Kalman filter the approach yields accurate training with the implication of accurate neural network model representing the input and output relationships in the application. As the process of concern is a dynamic system, the input source of information to neural network is time dependent so that the training algorithm presents an adaptive form for real-time operation for the monitoring task. (orig.)

Parallelization of Neural Network Training for NLP with Hogwild!

Directory of Open Access Journals (Sweden)

Deyringer Valentin

2017-10-01

Full Text Available Neural Networks are prevalent in todays NLP research. Despite their success for different tasks, training time is relatively long. We use Hogwild! to counteract this phenomenon and show that it is a suitable method to speed up training Neural Networks of different architectures and complexity. For POS tagging and translation we report considerable speedups of training, especially for the latter. We show that Hogwild! can be an important tool for training complex NLP architectures.

A simulation training evaluation method for distribution network fault based on radar chart

Directory of Open Access Journals (Sweden)

Yuhang Xu

2018-01-01

Full Text Available In order to solve the problem of automatic evaluation of dispatcher fault simulation training in distribution network, a simulation training evaluation method based on radar chart for distribution network fault is proposed. The fault handling information matrix is established to record the dispatcher fault handling operation sequence and operation information. The four situations of the dispatcher fault isolation operation are analyzed. The fault handling anti-misoperation rule set is established to describe the rules prohibiting dispatcher operation. Based on the idea of artificial intelligence reasoning, the feasibility of dispatcher fault handling is described by the feasibility index. The relevant factors and evaluation methods are discussed from the three aspects of the fault handling result feasibility, the anti-misoperation correctness and the operation process conciseness. The detailed calculation formula is given. Combining the independence and correlation between the three evaluation angles, a comprehensive evaluation method of distribution network fault simulation training based on radar chart is proposed. The method can comprehensively reflect the fault handling process of dispatchers, and comprehensively evaluate the fault handling process from various angles, which has good practical value.

Non-Linear State Estimation Using Pre-Trained Neural Networks

DEFF Research Database (Denmark)

Bayramoglu, Enis; Andersen, Nils Axel; Ravn, Ole

2010-01-01

effecting the transformation. This function is approximated by a neural network using offline training. The training is based on monte carlo sampling. A way to obtain parametric distributions of flexible shape to be used easily with these networks is also presented. The method can also be used to improve...... other parametric methods around regions with strong non-linearities by including them inside the network....

Influence of the Training Methods in the Diagnosis of Multiple Sclerosis Using Radial Basis Functions Artificial Neural Networks

Directory of Open Access Journals (Sweden)

Ángel Gutiérrez

2015-04-01

Full Text Available The data available in the average clinical study of a disease is very often small. This is one of the main obstacles in the application of neural networks to the classification of biological signals used for diagnosing diseases. A rule of thumb states that the number of parameters (weights that can be used for training a neural network should be around 15% of the available data, to avoid overlearning. This condition puts a limit on the dimension of the input space. Different authors have used different approaches to solve this problem, like eliminating redundancy in the data, preprocessing the data to find centers for the radial basis functions, or extracting a small number of features that were used as inputs. It is clear that the classification would be better the more features we could feed into the network. The approach utilized in this paper is incrementing the number of training elements with randomly expanding training sets. This way the number of original signals does not constraint the dimension of the input set in the radial basis network. Then we train the network using the method that minimizes the error function using the gradient descent algorithm and the method that uses the particle swarm optimization technique. A comparison between the two methods showed that for the same number of iterations on both methods, the particle swarm optimization was faster, it was learning to recognize only the sick people. On the other hand, the gradient method was not as good in general better at identifying those people.

The effects of working memory training on functional brain network efficiency.

Science.gov (United States)

Langer, Nicolas; von Bastian, Claudia C; Wirz, Helen; Oberauer, Klaus; Jäncke, Lutz

2013-10-01

The human brain is a highly interconnected network. Recent studies have shown that the functional and anatomical features of this network are organized in an efficient small-world manner that confers high efficiency of information processing at relatively low connection cost. However, it has been unclear how the architecture of functional brain networks is related to performance in working memory (WM) tasks and if these networks can be modified by WM training. Therefore, we conducted a double-blind training study enrolling 66 young adults. Half of the subjects practiced three WM tasks and were compared to an active control group practicing three tasks with low WM demand. High-density resting-state electroencephalography (EEG) was recorded before and after training to analyze graph-theoretical functional network characteristics at an intracortical level. WM performance was uniquely correlated with power in the theta frequency, and theta power was increased by WM training. Moreover, the better a person's WM performance, the more their network exhibited small-world topology. WM training shifted network characteristics in the direction of high performers, showing increased small-worldness within a distributed fronto-parietal network. Taken together, this is the first longitudinal study that provides evidence for the plasticity of the functional brain network underlying WM. Copyright © 2013 Elsevier Ltd. All rights reserved.

Training strategy for convolutional neural networks in pedestrian gender classification

Science.gov (United States)

Ng, Choon-Boon; Tay, Yong-Haur; Goi, Bok-Min

2017-06-01

In this work, we studied a strategy for training a convolutional neural network in pedestrian gender classification with limited amount of labeled training data. Unsupervised learning by k-means clustering on pedestrian images was used to learn the filters to initialize the first layer of the network. As a form of pre-training, supervised learning for the related task of pedestrian classification was performed. Finally, the network was fine-tuned for gender classification. We found that this strategy improved the network's generalization ability in gender classification, achieving better test results when compared to random weights initialization and slightly more beneficial than merely initializing the first layer filters by unsupervised learning. This shows that unsupervised learning followed by pre-training with pedestrian images is an effective strategy to learn useful features for pedestrian gender classification.

Behaviour in O of the Neural Networks Training Cost

DEFF Research Database (Denmark)

Goutte, Cyril

1998-01-01

We study the behaviour in zero of the derivatives of the cost function used when training non-linear neural networks. It is shown that a fair number offirst, second and higher order derivatives vanish in zero, validating the belief that 0 is a peculiar and potentially harmful location. These calc......We study the behaviour in zero of the derivatives of the cost function used when training non-linear neural networks. It is shown that a fair number offirst, second and higher order derivatives vanish in zero, validating the belief that 0 is a peculiar and potentially harmful location....... These calculations arerelated to practical and theoretical aspects of neural networks training....

Training Excitatory-Inhibitory Recurrent Neural Networks for Cognitive Tasks: A Simple and Flexible Framework.

Directory of Open Access Journals (Sweden)

H Francis Song

2016-02-01

Full Text Available The ability to simultaneously record from large numbers of neurons in behaving animals has ushered in a new era for the study of the neural circuit mechanisms underlying cognitive functions. One promising approach to uncovering the dynamical and computational principles governing population responses is to analyze model recurrent neural networks (RNNs that have been optimized to perform the same tasks as behaving animals. Because the optimization of network parameters specifies the desired output but not the manner in which to achieve this output, "trained" networks serve as a source of mechanistic hypotheses and a testing ground for data analyses that link neural computation to behavior. Complete access to the activity and connectivity of the circuit, and the ability to manipulate them arbitrarily, make trained networks a convenient proxy for biological circuits and a valuable platform for theoretical investigation. However, existing RNNs lack basic biological features such as the distinction between excitatory and inhibitory units (Dale's principle, which are essential if RNNs are to provide insights into the operation of biological circuits. Moreover, trained networks can achieve the same behavioral performance but differ substantially in their structure and dynamics, highlighting the need for a simple and flexible framework for the exploratory training of RNNs. Here, we describe a framework for gradient descent-based training of excitatory-inhibitory RNNs that can incorporate a variety of biological knowledge. We provide an implementation based on the machine learning library Theano, whose automatic differentiation capabilities facilitate modifications and extensions. We validate this framework by applying it to well-known experimental paradigms such as perceptual decision-making, context-dependent integration, multisensory integration, parametric working memory, and motor sequence generation. Our results demonstrate the wide range of neural

Nuclear safety education and training network

International Nuclear Information System (INIS)

Bastos, J.; Ulfkjaer, L.

2004-01-01

In March 2001, the Secretariat convened an Advisory Group on Education and Training in nuclear safety. The Advisory Group considered structure, scope and means related to the implementation of an IAEA Programme on Education and Training . A strategic plan was agreed and the following outputs were envisaged: 1. A Training Support Programme in nuclear safety, including a standardized and harmonized approach for training developed by the IAEA and in use by Member States. 2. National and regional training centres, established to support sustainable national nuclear safety infrastructures. 3. Training material for use by lecturers and students developed by the IAEA in English and translated to other languages. The implementation of the plan was initiated in 2002 emphasizing the preparation of training materials. In 2003 a pilot project for a network on Education and Training in Asia was initiated

Exploring empowerment in settings: mapping distributions of network power.

Science.gov (United States)

Neal, Jennifer Watling

2014-06-01

This paper brings together two trends in the empowerment literature-understanding empowerment in settings and understanding empowerment as relational-by examining what makes settings empowering from a social network perspective. Specifically, extending Neal and Neal's (Am J Community Psychol 48(3/4):157-167, 2011) conception of network power, an empowering setting is defined as one in which (1) actors have existing relationships that allow for the exchange of resources and (2) the distribution of network power among actors in the setting is roughly equal. The paper includes a description of how researchers can examine distributions of network power in settings. Next, this process is illustrated in both an abstract example and using empirical data on early adolescents' peer relationships in urban classrooms. Finally, implications for theory, methods, and intervention related to understanding empowering settings are explored.

Partially ordered sets in complex networks

International Nuclear Information System (INIS)

Xuan Qi; Du Fang; Wu Tiejun

2010-01-01

In this paper, a partial-order relation is defined among vertices of a network to describe which vertex is more important than another on its contribution to the connectivity of the network. A maximum linearly ordered subset of vertices is defined as a chain and the chains sharing the same end-vertex are grouped as a family. Through combining the same vertices appearing in different chains, a directed chain graph is obtained. Based on these definitions, a series of new network measurements, such as chain length distribution, family diversity distribution, as well as the centrality of families, are proposed. By studying the partially ordered sets in three kinds of real-world networks, many interesting results are revealed. For instance, the similar approximately power-law chain length distribution may be attributed to a chain-based positive feedback mechanism, i.e. new vertices prefer to participate in longer chains, which can be inferred by combining the notable preferential attachment rule with a well-ordered recommendation manner. Moreover, the relatively large average incoming degree of the chain graphs may indicate an efficient substitution mechanism in these networks. Most of the partially ordered set-based properties cannot be explained by the current well-known scale-free network models; therefore, we are required to propose more appropriate network models in the future.

A Comprehensive Training Data Set for the Development of Satellite-Based Volcanic Ash Detection Algorithms

Science.gov (United States)

Schmidl, Marius

2017-04-01

We present a comprehensive training data set covering a large range of atmospheric conditions, including disperse volcanic ash and desert dust layers. These data sets contain all information required for the development of volcanic ash detection algorithms based on artificial neural networks, urgently needed since volcanic ash in the airspace is a major concern of aviation safety authorities. Selected parts of the data are used to train the volcanic ash detection algorithm VADUGS. They contain atmospheric and surface-related quantities as well as the corresponding simulated satellite data for the channels in the infrared spectral range of the SEVIRI instrument on board MSG-2. To get realistic results, ECMWF, IASI-based, and GEOS-Chem data are used to calculate all parameters describing the environment, whereas the software package libRadtran is used to perform radiative transfer simulations returning the brightness temperatures for each atmospheric state. As optical properties are a prerequisite for radiative simulations accounting for aerosol layers, the development also included the computation of optical properties for a set of different aerosol types from different sources. A description of the developed software and the used methods is given, besides an overview of the resulting data sets.

Role of physical and mental training in brain network configuration.

Science.gov (United States)

Foster, Philip P

2015-01-01

It is hypothesized that the topology of brain networks is constructed by connecting nodes which may be continuously remodeled by appropriate training. Efficiency of physical and/or mental training on the brain relies on the flexibility of networks' architecture molded by local remodeling of proteins and synapses of excitatory neurons producing transformations in network topology. Continuous remodeling of proteins of excitatory neurons is fine-tuning the scaling and strength of excitatory synapses up or down via regulation of intra-cellular metabolic and regulatory networks of the genome-transcriptome-proteome interface. Alzheimer's disease is a model of "energy cost-driven small-world network disorder" with dysfunction of high-energy cost wiring as the network global efficiency is impaired by the deposition of an informed agent, the amyloid-β, selectively targeting high-degree nodes. In schizophrenia, the interconnectivity and density of rich-club networks are significantly reduced. Training-induced homeostatic synaptogenesis-enhancement, presumably via reconfiguration of brain networks into greater small-worldness, appears essential in learning, memory, and executive functions. A macroscopic cartography of creation-removal of synaptic connections in a macro-network, and at the intra-cellular scale, micro-networks regulate the physiological mechanisms for the preferential attachment of synapses. The strongest molecular relationship of exercise and functional connectivity was identified for brain-derived neurotrophic factor (BDNF). The allele variant, rs7294919, also shows a powerful relationship with the hippocampal volume. How the brain achieves this unique quest of reconfiguration remains a puzzle. What are the underlying mechanisms of synaptogenesis promoting communications brain ↔ muscle and brain ↔ brain in such trainings? What is the respective role of independent mental, physical, or combined-mental-physical trainings? Physical practice seems to be

Deliberate Practice of Creativity Training Set Series - A Creativity Training Material for Education (work in progress)

DEFF Research Database (Denmark)

Byrge, Christian

2018-01-01

Five training sets including 450 unique thinking direction cards and 120 exercise cards. Designed for Educational Purposes.......Five training sets including 450 unique thinking direction cards and 120 exercise cards. Designed for Educational Purposes....

Measuring dynamic process of working memory training with functional brain networks

Directory of Open Access Journals (Sweden)

Hong Wang

2015-12-01

Full Text Available In this paper, we proposed the functional brain networks and graphic theory method to measure the effect of working memory training on the neural activities. 12 subjects were recruited in this study, and they did the same working memory task before they had been trained and after training. We architected functional brain networks based on EEG coherence and calculated properties of brain networks to measure the neural co-activities and the working memory level of subjects. As the result, the internal connections in frontal region decreased after working memory training, but the connection between frontal region and top region increased. And the more small-world feature was observed after training. The features observed above were in alpha (8-13 Hz and beta (13-30 Hz bands. The functional brain networks based on EEG coherence proposed in this paper can be used as the indicator of working memory level.

EXPERIENCE NETWORKING UNIVERSITY OF EDUCATION TRAINING MASTERS SAFETY OF LIFE

Directory of Open Access Journals (Sweden)

Elvira Mikhailovna Rebko

2016-02-01

Full Text Available The article discloses experience networking of universities (Herzen State Pedagogical University and Sakhalin State University in the development and implementation of joint training programs for master’s education in the field of life safety «Social security in the urban environment». The novelty of the work is to create a schematic design of basic educational training program for master’s education in the mode of networking, and to identify effective instructional techniques and conditions of networking.Purpose – present the results of the joint development of a network of the basic educational program (BEP, to identify the stages of networking, to design a generalized scheme of development and implementation of a network of educational training program for master’s education in the field of life safety.Results generalized model of networking partner institutions to develop and implement the basic educational program master.Practical implications: the education process for Master of Education in the field of health and safety in Herzen State Pedagogical University and Sakhalin State University.

Combined Ozone Retrieval From METOP Sensors Using META-Training Of Deep Neural Networks

Science.gov (United States)

Felder, Martin; Sehnke, Frank; Kaifel, Anton

2013-12-01

The newest installment of our well-proven Neural Net- work Ozone Retrieval System (NNORSY) combines the METOP sensors GOME-2 and IASI with cloud information from AVHRR. Through the use of advanced meta- learning techniques like automatic feature selection and automatic architecture search applied to a set of deep neural networks, having at least two or three hidden layers, we have been able to avoid many technical issues normally encountered during the construction of such a joint retrieval system. This has been made possible by harnessing the processing power of modern consumer graphics cards with high performance graphic processors (GPU), which decreases training times by about two orders of magnitude. The system was trained on data from 2009 and 2010, including target ozone profiles from ozone sondes, ACE- FTS and MLS-AURA. To make maximum use of tropospheric information in the spectra, the data were partitioned into several sets of different cloud fraction ranges with the GOME-2 FOV, on which specialized retrieval networks are being trained. For the final ozone retrieval processing the different specialized networks are combined. The resulting retrieval system is very stable and does not show any systematic dependence on solar zenith angle, scan angle or sensor degradation. We present several sensitivity studies with regard to cloud fraction and target sensor type, as well as the performance in several latitude bands and with respect to independent validation stations. A visual cross-comparison against high-resolution ozone profiles from the KNMI EUMETSAT Ozone SAF product has also been performed and shows some distinctive features which we will briefly discuss. Overall, we demonstrate that a complex retrieval system can now be constructed with a minimum of ma- chine learning knowledge, using automated algorithms for many design decisions previously requiring expert knowledge. Provided sufficient training data and computation power of GPUs is available, the

C-RNN-GAN: Continuous recurrent neural networks with adversarial training

OpenAIRE

Mogren, Olof

2016-01-01

Generative adversarial networks have been proposed as a way of efficiently training deep generative neural networks. We propose a generative adversarial model that works on continuous sequential data, and apply it by training it on a collection of classical music. We conclude that it generates music that sounds better and better as the model is trained, report statistics on generated music, and let the reader judge the quality by downloading the generated songs.

Training Excitatory-Inhibitory Recurrent Neural Networks for Cognitive Tasks: A Simple and Flexible Framework

Science.gov (United States)

Wang, Xiao-Jing

2016-01-01

The ability to simultaneously record from large numbers of neurons in behaving animals has ushered in a new era for the study of the neural circuit mechanisms underlying cognitive functions. One promising approach to uncovering the dynamical and computational principles governing population responses is to analyze model recurrent neural networks (RNNs) that have been optimized to perform the same tasks as behaving animals. Because the optimization of network parameters specifies the desired output but not the manner in which to achieve this output, “trained” networks serve as a source of mechanistic hypotheses and a testing ground for data analyses that link neural computation to behavior. Complete access to the activity and connectivity of the circuit, and the ability to manipulate them arbitrarily, make trained networks a convenient proxy for biological circuits and a valuable platform for theoretical investigation. However, existing RNNs lack basic biological features such as the distinction between excitatory and inhibitory units (Dale’s principle), which are essential if RNNs are to provide insights into the operation of biological circuits. Moreover, trained networks can achieve the same behavioral performance but differ substantially in their structure and dynamics, highlighting the need for a simple and flexible framework for the exploratory training of RNNs. Here, we describe a framework for gradient descent-based training of excitatory-inhibitory RNNs that can incorporate a variety of biological knowledge. We provide an implementation based on the machine learning library Theano, whose automatic differentiation capabilities facilitate modifications and extensions. We validate this framework by applying it to well-known experimental paradigms such as perceptual decision-making, context-dependent integration, multisensory integration, parametric working memory, and motor sequence generation. Our results demonstrate the wide range of neural activity

Mining big data sets of plankton images: a zero-shot learning approach to retrieve labels without training data

Science.gov (United States)

Orenstein, E. C.; Morgado, P. M.; Peacock, E.; Sosik, H. M.; Jaffe, J. S.

2016-02-01

Technological advances in instrumentation and computing have allowed oceanographers to develop imaging systems capable of collecting extremely large data sets. With the advent of in situ plankton imaging systems, scientists must now commonly deal with "big data" sets containing tens of millions of samples spanning hundreds of classes, making manual classification untenable. Automated annotation methods are now considered to be the bottleneck between collection and interpretation. Typically, such classifiers learn to approximate a function that predicts a predefined set of classes for which a considerable amount of labeled training data is available. The requirement that the training data span all the classes of concern is problematic for plankton imaging systems since they sample such diverse, rapidly changing populations. These data sets may contain relatively rare, sparsely distributed, taxa that will not have associated training data; a classifier trained on a limited set of classes will miss these samples. The computer vision community, leveraging advances in Convolutional Neural Networks (CNNs), has recently attempted to tackle such problems using "zero-shot" object categorization methods. Under a zero-shot framework, a classifier is trained to map samples onto a set of attributes rather than a class label. These attributes can include visual and non-visual information such as what an organism is made out of, where it is distributed globally, or how it reproduces. A second stage classifier is then used to extrapolate a class. In this work, we demonstrate a zero-shot classifier, implemented with a CNN, to retrieve out-of-training-set labels from images. This method is applied to data from two continuously imaging, moored instruments: the Scripps Plankton Camera System (SPCS) and the Imaging FlowCytobot (IFCB). Results from simulated deployment scenarios indicate zero-shot classifiers could be successful at recovering samples of rare taxa in image sets. This

The Analysis of User Behaviour of a Network Management Training Tool using a Neural Network

Directory of Open Access Journals (Sweden)

Helen Donelan

2005-10-01

Full Text Available A novel method for the analysis and interpretation of data that describes the interaction between trainee network managers and a network management training tool is presented. A simulation based approach is currently being used to train network managers, through the use of a simulated network. The motivation is to provide a tool for exposing trainees to a life like situation without disrupting a live network. The data logged by this system describes the detailed interaction between trainee network manager and simulated network. The work presented here provides an analysis of this interaction data that enables an assessment of the capabilities of the trainee network manager as well as an understanding of how the network management tasks are being approached. A neural network architecture is implemented in order to perform an exploratory data analysis of the interaction data. The neural network employs a novel form of continuous self-organisation to discover key features in the data and thus provide new insights into the learning and teaching strategies employed.

Method Accelerates Training Of Some Neural Networks

Science.gov (United States)

Shelton, Robert O.

1992-01-01

Three-layer networks trained faster provided two conditions are satisfied: numbers of neurons in layers are such that majority of work done in synaptic connections between input and hidden layers, and number of neurons in input layer at least as great as number of training pairs of input and output vectors. Based on modified version of back-propagation method.

Character Recognition Using Genetically Trained Neural Networks

Energy Technology Data Exchange (ETDEWEB)

Diniz, C.; Stantz, K.M.; Trahan, M.W.; Wagner, J.S.

1998-10-01

Computationally intelligent recognition of characters and symbols addresses a wide range of applications including foreign language translation and chemical formula identification. The combination of intelligent learning and optimization algorithms with layered neural structures offers powerful techniques for character recognition. These techniques were originally developed by Sandia National Laboratories for pattern and spectral analysis; however, their ability to optimize vast amounts of data make them ideal for character recognition. An adaptation of the Neural Network Designer soflsvare allows the user to create a neural network (NN_) trained by a genetic algorithm (GA) that correctly identifies multiple distinct characters. The initial successfid recognition of standard capital letters can be expanded to include chemical and mathematical symbols and alphabets of foreign languages, especially Arabic and Chinese. The FIN model constructed for this project uses a three layer feed-forward architecture. To facilitate the input of characters and symbols, a graphic user interface (GUI) has been developed to convert the traditional representation of each character or symbol to a bitmap. The 8 x 8 bitmap representations used for these tests are mapped onto the input nodes of the feed-forward neural network (FFNN) in a one-to-one correspondence. The input nodes feed forward into a hidden layer, and the hidden layer feeds into five output nodes correlated to possible character outcomes. During the training period the GA optimizes the weights of the NN until it can successfully recognize distinct characters. Systematic deviations from the base design test the network's range of applicability. Increasing capacity, the number of letters to be recognized, requires a nonlinear increase in the number of hidden layer neurodes. Optimal character recognition performance necessitates a minimum threshold for the number of cases when genetically training the net. And, the

Dynamic training algorithm for dynamic neural networks

International Nuclear Information System (INIS)

Tan, Y.; Van Cauwenberghe, A.; Liu, Z.

1996-01-01

The widely used backpropagation algorithm for training neural networks based on the gradient descent has a significant drawback of slow convergence. A Gauss-Newton method based recursive least squares (RLS) type algorithm with dynamic error backpropagation is presented to speed-up the learning procedure of neural networks with local recurrent terms. Finally, simulation examples concerning the applications of the RLS type algorithm to identification of nonlinear processes using a local recurrent neural network are also included in this paper

Training and validation of the ATLAS pixel clustering neural networks

CERN Document Server

The ATLAS collaboration

2018-01-01

The high centre-of-mass energy of the LHC gives rise to dense environments, such as the core of high-pT jets, in which the charge clusters left by ionising particles in the silicon sensors of the pixel detector can merge, compromising the tracking and vertexing efficiency. To recover optimal performance, a neural network-based approach is used to separate clusters originating from single and multiple particles and to estimate all hit positions within clusters. This note presents the training strategy employed and a set of benchmark performance measurements on a Monte Carlo sample of high-pT dijet events.

Accelerating deep neural network training with inconsistent stochastic gradient descent.

Science.gov (United States)

Wang, Linnan; Yang, Yi; Min, Renqiang; Chakradhar, Srimat

2017-09-01

Stochastic Gradient Descent (SGD) updates Convolutional Neural Network (CNN) with a noisy gradient computed from a random batch, and each batch evenly updates the network once in an epoch. This model applies the same training effort to each batch, but it overlooks the fact that the gradient variance, induced by Sampling Bias and Intrinsic Image Difference, renders different training dynamics on batches. In this paper, we develop a new training strategy for SGD, referred to as Inconsistent Stochastic Gradient Descent (ISGD) to address this problem. The core concept of ISGD is the inconsistent training, which dynamically adjusts the training effort w.r.t the loss. ISGD models the training as a stochastic process that gradually reduces down the mean of batch's loss, and it utilizes a dynamic upper control limit to identify a large loss batch on the fly. ISGD stays on the identified batch to accelerate the training with additional gradient updates, and it also has a constraint to penalize drastic parameter changes. ISGD is straightforward, computationally efficient and without requiring auxiliary memories. A series of empirical evaluations on real world datasets and networks demonstrate the promising performance of inconsistent training. Copyright © 2017 Elsevier Ltd. All rights reserved.

Case Study: Does training of private networks of Family Planning clinicians in urban Pakistan affect service utilization?

Science.gov (United States)

2010-01-01

Background To determine whether training of providers participating in franchise clinic networks is associated with increased Family Planning service use among low-income urban families in Pakistan. Methods The study uses 2001 survey data consisting of interviews with 1113 clinical and non-clinical providers working in public and private hospitals/clinics. Data analysis excludes non-clinical providers reducing sample size to 822. Variables for the analysis are divided into client volume, and training in family planning. Regression models are used to compute the association between training and service use in franchise versus private non-franchise clinics. Results In franchise clinic networks, staff are 6.5 times more likely to receive family planning training (P = 0.00) relative to private non-franchises. Service use was significantly associated with training (P = 0.00), franchise affiliation (P = 0.01), providers' years of family planning experience (P = 0.02) and the number of trained staff working at government owned clinics (P = 0.00). In this setting, nurses are significantly less likely to receive training compared to doctors (P = 0.00). Conclusions These findings suggest that franchises recruit and train various cadres of health workers and training maybe associated with increased service use through improvement in quality of services. PMID:21062460

Case Study: Does training of private networks of Family Planning clinicians in urban Pakistan affect service utilization?

Directory of Open Access Journals (Sweden)

Qureshi Asma M

2010-11-01

Full Text Available Abstract Background To determine whether training of providers participating in franchise clinic networks is associated with increased Family Planning service use among low-income urban families in Pakistan. Methods The study uses 2001 survey data consisting of interviews with 1113 clinical and non-clinical providers working in public and private hospitals/clinics. Data analysis excludes non-clinical providers reducing sample size to 822. Variables for the analysis are divided into client volume, and training in family planning. Regression models are used to compute the association between training and service use in franchise versus private non-franchise clinics. Results In franchise clinic networks, staff are 6.5 times more likely to receive family planning training (P = 0.00 relative to private non-franchises. Service use was significantly associated with training (P = 0.00, franchise affiliation (P = 0.01, providers' years of family planning experience (P = 0.02 and the number of trained staff working at government owned clinics (P = 0.00. In this setting, nurses are significantly less likely to receive training compared to doctors (P = 0.00. Conclusions These findings suggest that franchises recruit and train various cadres of health workers and training maybe associated with increased service use through improvement in quality of services.

Case Study: Does training of private networks of Family Planning clinicians in urban Pakistan affect service utilization?

Science.gov (United States)

Qureshi, Asma M

2010-11-09

To determine whether training of providers participating in franchise clinic networks is associated with increased Family Planning service use among low-income urban families in Pakistan. The study uses 2001 survey data consisting of interviews with 1113 clinical and non-clinical providers working in public and private hospitals/clinics. Data analysis excludes non-clinical providers reducing sample size to 822. Variables for the analysis are divided into client volume, and training in family planning. Regression models are used to compute the association between training and service use in franchise versus private non-franchise clinics. In franchise clinic networks, staff are 6.5 times more likely to receive family planning training (P = 0.00) relative to private non-franchises. Service use was significantly associated with training (P = 0.00), franchise affiliation (P = 0.01), providers' years of family planning experience (P = 0.02) and the number of trained staff working at government owned clinics (P = 0.00). In this setting, nurses are significantly less likely to receive training compared to doctors (P = 0.00). These findings suggest that franchises recruit and train various cadres of health workers and training maybe associated with increased service use through improvement in quality of services.

Gradual DropIn of Layers to Train Very Deep Neural Networks

OpenAIRE

Smith, Leslie N.; Hand, Emily M.; Doster, Timothy

2015-01-01

We introduce the concept of dynamically growing a neural network during training. In particular, an untrainable deep network starts as a trainable shallow network and newly added layers are slowly, organically added during training, thereby increasing the network's depth. This is accomplished by a new layer, which we call DropIn. The DropIn layer starts by passing the output from a previous layer (effectively skipping over the newly added layers), then increasingly including units from the ne...

Method of Parallel-Hierarchical Network Self-Training and its Application for Pattern Classification and Recognition

Directory of Open Access Journals (Sweden)

TIMCHENKO, L.

2012-11-01

Full Text Available Propositions necessary for development of parallel-hierarchical (PH network training methods are discussed in this article. Unlike already known structures of the artificial neural network, where non-normalized (absolute similarity criteria are used for comparison, the suggested structure uses a normalized criterion. Based on the analysis of training rules, a conclusion is made that application of two training methods with a teacher is optimal for PH network training: error correction-based training and memory-based training. Mathematical models of training and a combined method of PH network training for recognition of static and dynamic patterns are developed.

Jointly Optimize Data Augmentation and Network Training: Adversarial Data Augmentation in Human Pose Estimation

OpenAIRE

Peng, Xi; Tang, Zhiqiang; Yang, Fei; Feris, Rogerio; Metaxas, Dimitris

2018-01-01

Random data augmentation is a critical technique to avoid overfitting in training deep neural network models. However, data augmentation and network training are usually treated as two isolated processes, limiting the effectiveness of network training. Why not jointly optimize the two? We propose adversarial data augmentation to address this limitation. The main idea is to design an augmentation network (generator) that competes against a target network (discriminator) by generating `hard' au...

Role of physical and mental training in brain network configuration

Directory of Open Access Journals (Sweden)

Philip P. Foster

2015-06-01

Full Text Available Continuous remodeling of proteins of excitatory neurons is fine-tuning the scaling and strength of excitatory synapses up or down via regulation of intra-cellular metabolic and regulatory networks of the genome-transcriptome-proteome interface. Alzheimer's disease is a model of energy cost-driven small-world network disorder as the network global efficiency is impaired by the deposition of an informed agent, the amyloid-β, selectively targeting high-degree nodes. In schizophrenia, the interconnectivity and density of rich-club networks are significantly reduced. Training-induced homeostatic synaptogenesis-enhancement produces a reconfiguration of brain networks into greater small-worldness. Creation of synaptic connections in a macro-network, and, at the intra-cellular scale, micro-networks regulate the physiological mechanisms for the preferential attachment of synapses. The strongest molecular relationship of exercise and functional connectivity was identified for brain-derived neurotrophic factor (BDNF. The allele variant, rs7294919, also shows a powerful relationship with the hippocampal volume. How the brain achieves this unique quest of reconfiguration remains a puzzle. What are the underlying mechanisms of synaptogenesis promoting communications brain ↔ muscle and brain ↔ brain in such trainings? What is the respective role of independent mental, physical or combined-mental-physical trainings? Physical practice seems to be playing an instrumental role in the cognitive enhancement (brain ↔ muscle com.. However, mental training, meditation or virtual reality (films, games require only minimal motor activity and cardio-respiratory stimulation. Therefore, other potential paths (brain ↔ brain com. molding brain networks are nonetheless essential. Patients with motor neuron disease/injury (e.g. amyotrophic lateral sclerosis, traumatism also achieve successful cognitive enhancement albeit they may only elicit mental practice

Working memory training mostly engages general-purpose large-scale networks for learning.

Science.gov (United States)

Salmi, Juha; Nyberg, Lars; Laine, Matti

2018-03-21

The present meta-analytic study examined brain activation changes following working memory (WM) training, a form of cognitive training that has attracted considerable interest. Comparisons with perceptual-motor (PM) learning revealed that WM training engages domain-general large-scale networks for learning encompassing the dorsal attention and salience networks, sensory areas, and striatum. Also the dynamics of the training-induced brain activation changes within these networks showed a high overlap between WM and PM training. The distinguishing feature for WM training was the consistent modulation of the dorso- and ventrolateral prefrontal cortex (DLPFC/VLPFC) activity. The strongest candidate for mediating transfer to similar untrained WM tasks was the frontostriatal system, showing higher striatal and VLPFC activations, and lower DLPFC activations after training. Modulation of transfer-related areas occurred mostly with longer training periods. Overall, our findings place WM training effects into a general perception-action cycle, where some modulations may depend on the specific cognitive demands of a training task. Copyright © 2018 Elsevier Ltd. All rights reserved.

Network Training for a Boy with Learning Disabilities and Behaviours That Challenge

Science.gov (United States)

Cooper, Kate; McElwee, Jennifer

2016-01-01

Background: Network Training is an intervention that draws upon systemic ideas and behavioural principles to promote positive change in networks of support for people defined as having a learning disability. To date, there are no published case studies looking at the outcomes of Network Training. Materials and Methods: This study aimed to…

Maximizing lipocalin prediction through balanced and diversified training set and decision fusion.

Science.gov (United States)

Nath, Abhigyan; Subbiah, Karthikeyan

2015-12-01

Lipocalins are short in sequence length and perform several important biological functions. These proteins are having less than 20% sequence similarity among paralogs. Experimentally identifying them is an expensive and time consuming process. The computational methods based on the sequence similarity for allocating putative members to this family are also far elusive due to the low sequence similarity existing among the members of this family. Consequently, the machine learning methods become a viable alternative for their prediction by using the underlying sequence/structurally derived features as the input. Ideally, any machine learning based prediction method must be trained with all possible variations in the input feature vector (all the sub-class input patterns) to achieve perfect learning. A near perfect learning can be achieved by training the model with diverse types of input instances belonging to the different regions of the entire input space. Furthermore, the prediction performance can be improved through balancing the training set as the imbalanced data sets will tend to produce the prediction bias towards majority class and its sub-classes. This paper is aimed to achieve (i) the high generalization ability without any classification bias through the diversified and balanced training sets as well as (ii) enhanced the prediction accuracy by combining the results of individual classifiers with an appropriate fusion scheme. Instead of creating the training set randomly, we have first used the unsupervised Kmeans clustering algorithm to create diversified clusters of input patterns and created the diversified and balanced training set by selecting an equal number of patterns from each of these clusters. Finally, probability based classifier fusion scheme was applied on boosted random forest algorithm (which produced greater sensitivity) and K nearest neighbour algorithm (which produced greater specificity) to achieve the enhanced predictive performance

Smooth function approximation using neural networks.

Science.gov (United States)

Ferrari, Silvia; Stengel, Robert F

2005-01-01

An algebraic approach for representing multidimensional nonlinear functions by feedforward neural networks is presented. In this paper, the approach is implemented for the approximation of smooth batch data containing the function's input, output, and possibly, gradient information. The training set is associated to the network adjustable parameters by nonlinear weight equations. The cascade structure of these equations reveals that they can be treated as sets of linear systems. Hence, the training process and the network approximation properties can be investigated via linear algebra. Four algorithms are developed to achieve exact or approximate matching of input-output and/or gradient-based training sets. Their application to the design of forward and feedback neurocontrollers shows that algebraic training is characterized by faster execution speeds and better generalization properties than contemporary optimization techniques.

A neural network driving curve generation method for the heavy-haul train

Directory of Open Access Journals (Sweden)

Youneng Huang

2016-05-01

Full Text Available The heavy-haul train has a series of characteristics, such as the locomotive traction properties, the longer length of train, and the nonlinear train pipe pressure during train braking. When the train is running on a continuous long and steep downgrade railway line, the safety of the train is ensured by cycle braking, which puts high demands on the driving skills of the driver. In this article, a driving curve generation method for the heavy-haul train based on a neural network is proposed. First, in order to describe the nonlinear characteristics of train braking, the neural network model is constructed and trained by practical driving data. In the neural network model, various nonlinear neurons are interconnected to work for information processing and transmission. The target value of train braking pressure reduction and release time is achieved by modeling the braking process. The equation of train motion is computed to obtain the driving curve. Finally, in four typical operation scenarios, comparing the curve data generated by the method with corresponding practical data of the Shuohuang heavy-haul railway line, the results show that the method is effective.

Contemporary social network sites: Relevance in anesthesiology teaching, training, and research.

Science.gov (United States)

Haldar, Rudrashish; Kaushal, Ashutosh; Samanta, Sukhen; Ambesh, Paurush; Srivastava, Shashi; Singh, Prabhat K

2016-01-01

The phenomenal popularity of social networking sites has been used globally by medical professionals to boost professional associations and scientific developments. They have tremendous potential to forge professional liaisons, generate employment,upgrading skills and publicizing scientific achievements. We highlight the role of social networking mediums in influencing teaching, training and research in anaesthesiology. The growth of social networking sites have been prompted by the limitations of previous facilities in terms of ease of data and interface sharing and the amalgamation of audio visual aids on common platforms in the newer facilities. Contemporary social networking sites like Facebook, Twitter, Tumblr,Linkedn etc and their respective features based on anaesthesiology training or practice have been discussed. A host of advantages which these sites confer are also discussed. Likewise the potential pitfalls and drawbacks of these facilities have also been addressed. Social networking sites have immense potential for development of training and research in Anaesthesiology. However responsible and cautious utilization is advocated.

Ebola 2014: Setting up a port health screening programme at an international train station.

Science.gov (United States)

Cleary, Vivien; Wynne-Evans, Edward; Freed, James; Fleet, Katie; Thorn, Simone; Turbitt, Deborah

2017-12-01

An outbreak of Ebola virus disease (EVD) began in Guinea in December 2013 and was declared a Public Health Emergency of International Concern by the World Health Organization in August 2014. In October, the UK government tasked Public Health England (PHE) to set up EVD screening at key ports. The key aim of port-of-entry screening was to identify passengers coming from areas with high risk of EVD, and give them advice to raise their awareness of symptoms and what actions to take. Direct flights from Sierra Leone, Guinea or Liberia had all been cancelled, so intelligence on passenger numbers and routes was used to identify the most commonly used routes from the affected countries into the UK. One of these was St Pancras International train station. Screening had never previously been implemented at a UK train station so had to be set up from scratch. Key to the success of this was excellent multi-agency working between PHE, the UK Border Force, Eurostar, Network Rail and the Cabinet Office. This paper gives an overview of the activation of EVD screening at St Pancras International and the subsequent decommissioning.

STACCATO: a novel solution to supernova photometric classification with biased training sets

Science.gov (United States)

Revsbech, E. A.; Trotta, R.; van Dyk, D. A.

2018-01-01

We present a new solution to the problem of classifying Type Ia supernovae from their light curves alone given a spectroscopically confirmed but biased training set, circumventing the need to obtain an observationally expensive unbiased training set. We use Gaussian processes (GPs) to model the supernovae's (SN's) light curves, and demonstrate that the choice of covariance function has only a small influence on the GPs ability to accurately classify SNe. We extend and improve the approach of Richards et al. - a diffusion map combined with a random forest classifier - to deal specifically with the case of biased training sets. We propose a novel method called Synthetically Augmented Light Curve Classification (STACCATO) that synthetically augments a biased training set by generating additional training data from the fitted GPs. Key to the success of the method is the partitioning of the observations into subgroups based on their propensity score of being included in the training set. Using simulated light curve data, we show that STACCATO increases performance, as measured by the area under the Receiver Operating Characteristic curve (AUC), from 0.93 to 0.96, close to the AUC of 0.977 obtained using the 'gold standard' of an unbiased training set and significantly improving on the previous best result of 0.88. STACCATO also increases the true positive rate for SNIa classification by up to a factor of 50 for high-redshift/low-brightness SNe.

Autocatalytic sets in a partitioned biochemical network.

Science.gov (United States)

Smith, Joshua I; Steel, Mike; Hordijk, Wim

2014-01-01

In previous work, RAF theory has been developed as a tool for making theoretical progress on the origin of life question, providing insight into the structure and occurrence of self-sustaining and collectively autocatalytic sets within catalytic polymer networks. We present here an extension in which there are two "independent" polymer sets, where catalysis occurs within and between the sets, but there are no reactions combining polymers from both sets. Such an extension reflects the interaction between nucleic acids and peptides observed in modern cells and proposed forms of early life. We present theoretical work and simulations which suggest that the occurrence of autocatalytic sets is robust to the partitioned structure of the network. We also show that autocatalytic sets remain likely even when the molecules in the system are not polymers, and a low level of inhibition is present. Finally, we present a kinetic extension which assigns a rate to each reaction in the system, and show that identifying autocatalytic sets within such a system is an NP-complete problem. Recent experimental work has challenged the necessity of an RNA world by suggesting that peptide-nucleic acid interactions occurred early in chemical evolution. The present work indicates that such a peptide-RNA world could support the spontaneous development of autocatalytic sets and is thus a feasible alternative worthy of investigation.

European training network on full-parallax imaging (Conference Presentation)

Science.gov (United States)

Martínez-Corral, Manuel; Saavedra, Genaro

2017-05-01

Current displays are far from truly recreating visual reality. This requires a full-parallax display that can reproduce radiance field emanated from the real scenes. The develop-ment of such technology will require a new generation of researchers trained both in the physics, and in the biology of human vision. The European Training Network on Full-Parallax Imaging (ETN-FPI) aims at developing this new generation. Under H2020 funding ETN-FPI brings together 8 beneficiaries and 8 partner organizations from five EU countries with the aim of training 15 talented pre-doctoral students to become future research leaders in this area. In this contribution we will explain the main objectives of the network, and specifically the advances obtained at the University of Valencia.

EXPERIENCE NETWORKING UNIVERSITY OF EDUCATION TRAINING MASTERS SAFETY OF LIFE

OpenAIRE

Elvira Mikhailovna Rebko

2016-01-01

The article discloses experience networking of universities (Herzen State Pedagogical University and Sakhalin State University) in the development and implementation of joint training programs for master’s education in the field of life safety «Social security in the urban environment». The novelty of the work is to create a schematic design of basic educational training program for master’s education in the mode of networking, and to identify effective instructional techniques and conditions...

NeuroRecovery Network provides standardization of locomotor training for persons with incomplete spinal cord injury.

Science.gov (United States)

Morrison, Sarah A; Forrest, Gail F; VanHiel, Leslie R; Davé, Michele; D'Urso, Denise

2012-09-01

To illustrate the continuity of care afforded by a standardized locomotor training program across a multisite network setting within the Christopher and Dana Reeve Foundation NeuroRecovery Network (NRN). Single patient case study. Two geographically different hospital-based outpatient facilities. This case highlights a 25-year-old man diagnosed with C4 motor incomplete spinal cord injury with American Spinal Injury Association Impairment Scale grade D. Standardized locomotor training program 5 sessions per week for 1.5 hours per session, for a total of 100 treatment sessions, with 40 sessions at 1 center and 60 at another. Ten-meter walk test and 6-minute walk test were assessed at admission and discharge across both facilities. For each of the 100 treatment sessions percent body weight support, average, and maximum treadmill speed were evaluated. Locomotor endurance, as measured by the 6-minute walk test, and overground gait speed showed consistent improvement from admission to discharge. Throughout training, the patient decreased the need for body weight support and was able to tolerate faster treadmill speeds. Data indicate that the patient continued to improve on both treatment parameters and walking function. Standardization across the NRN centers provided a mechanism for delivering consistent and reproducible locomotor training programs across 2 facilities without disrupting training or recovery progression. Copyright © 2012 American Congress of Rehabilitation Medicine. Published by Elsevier Inc. All rights reserved.

The lateralization of intrinsic networks in the aging brain implicates the effects of cognitive training

Directory of Open Access Journals (Sweden)

Cheng eLuo

2016-03-01

Full Text Available Lateralization of function is an important organization of human brain. The distribution of intrinsic networks in the resting brain is strongly related to the cognitive function, gender and age. In this study, the longitudinal design with one year duration was used to evaluate the cognitive training effects on the lateralization of intrinsic networks among healthy older adults. The subjects were divided into two groups randomly: one with multi-domain cognitive training in three month, the other as a wait-list control group. Resting state fMRI data were acquired before training and one year after training. We analyzed the functional lateralization in ten common resting state fMRI networks. We observed statically significant training effects on the lateralization of two important RSNs related to high-level cognition: right- and left- frontoparietal networks. Especially, the lateralization of left-frontoparietal network were retained well in training group, but decreased in control group. The increased lateralization with aging was observed on the cerebellum network, in which the lateralization was significantly increased in control group although the same change tendency was observed in training group. These findings indicate that the lateralization of the high-level cognitive intrinsic networks is sensitive to the multi-domain cognitive training. This study provides a neuroimaging evidence to support that the cognitive training should have advantages to the cognitive decline in healthy older adults.

Training set extension for SVM ensemble in P300-speller with familiar face paradigm.

Science.gov (United States)

Li, Qi; Shi, Kaiyang; Gao, Ning; Li, Jian; Bai, Ou

2018-03-27

P300-spellers are brain-computer interface (BCI)-based character input systems. Support vector machine (SVM) ensembles are trained with large-scale training sets and used as classifiers in these systems. However, the required large-scale training data necessitate a prolonged collection time for each subject, which results in data collected toward the end of the period being contaminated by the subject's fatigue. This study aimed to develop a method for acquiring more training data based on a collected small training set. A new method was developed in which two corresponding training datasets in two sequences are superposed and averaged to extend the training set. The proposed method was tested offline on a P300-speller with the familiar face paradigm. The SVM ensemble with extended training set achieved 85% classification accuracy for the averaged results of four sequences, and 100% for 11 sequences in the P300-speller. In contrast, the conventional SVM ensemble with non-extended training set achieved only 65% accuracy for four sequences, and 92% for 11 sequences. The SVM ensemble with extended training set achieves higher classification accuracies than the conventional SVM ensemble, which verifies that the proposed method effectively improves the classification performance of BCI P300-spellers, thus enhancing their practicality.

Estimation of Collapse Moment for Wall Thinned Elbows Using Fuzzy Neural Networks

International Nuclear Information System (INIS)

Na, Man Gyun; Kim, Jin Weon; Shin, Sun Ho; Kim, Koung Suk; Kang, Ki Soo

2004-01-01

In this work, the collapse moment due to wall-thinning defects is estimated by using fuzzy neural networks. The developed fuzzy neural networks have been applied to the numerical data obtained from the finite element analysis. Principal component analysis is used to preprocess the input signals into the fuzzy neural network to reduce the sensitivity to the input change and the fuzzy neural networks are trained by using the data set prepared for training (training data) and verified by using another data set different (independent) from the training data. Also, two fuzzy neural networks are trained for two data sets divided into the two classes of extrados and intrados defects, which is because they have different characteristics. The relative 2-sigma errors of the estimated collapse moment are 3.07% for the training data and 4.12% for the test data. It is known from this result that the fuzzy neural networks are sufficiently accurate to be used in the wall-thinning monitoring of elbows

Solar Training Network and Solar Ready Vets

Energy Technology Data Exchange (ETDEWEB)

Dalstrom, Tenley Ann

2016-09-14

In 2016, the White House announced the Solar Ready Vets program, funded under DOE's SunShot initiative would be administered by The Solar Foundation to connect transitioning military personnel to solar training and employment as they separate from service. This presentation is geared to informing and recruiting employer partners for the Solar Ready Vets program, and the Solar Training Network. It describes the programs, and the benefits to employers that choose to connect to the programs.

Train-Network Interactions and Stability Evaluation in High-Speed Railways--Part I: Phenomena and Modeling

DEFF Research Database (Denmark)

Hu, Haitao; Tao, Haidong; Blaabjerg, Frede

2018-01-01

of the electric trains and traction network are equally modeled. In which, an impedance-based input behavior of the train is fully investigated with considering available controllers and their parameters in DQ-domain. While, the entire traction network, including traction transformer, catenary, supply lines......, is represented in a frequency-domain nodal matrix. Furthermore, the impedance-frequency responses of both electric train and traction network are measured and validated through frequency scan method. Finally, a generalized train-network simulation and experimental systems are proposed for verifying...

Neural-Network Object-Recognition Program

Science.gov (United States)

Spirkovska, L.; Reid, M. B.

1993-01-01

HONTIOR computer program implements third-order neural network exhibiting invariance under translation, change of scale, and in-plane rotation. Invariance incorporated directly into architecture of network. Only one view of each object needed to train network for two-dimensional-translation-invariant recognition of object. Also used for three-dimensional-transformation-invariant recognition by training network on only set of out-of-plane rotated views. Written in C language.

Training feed-forward neural networks with gain constraints

Science.gov (United States)

Hartman

2000-04-01

Inaccurate input-output gains (partial derivatives of outputs with respect to inputs) are common in neural network models when input variables are correlated or when data are incomplete or inaccurate. Accurate gains are essential for optimization, control, and other purposes. We develop and explore a method for training feedforward neural networks subject to inequality or equality-bound constraints on the gains of the learned mapping. Gain constraints are implemented as penalty terms added to the objective function, and training is done using gradient descent. Adaptive and robust procedures are devised for balancing the relative strengths of the various terms in the objective function, which is essential when the constraints are inconsistent with the data. The approach has the virtue that the model domain of validity can be extended via extrapolation training, which can dramatically improve generalization. The algorithm is demonstrated here on artificial and real-world problems with very good results and has been advantageously applied to dozens of models currently in commercial use.

Efficient probabilistic inference in generic neural networks trained with non-probabilistic feedback.

Science.gov (United States)

Orhan, A Emin; Ma, Wei Ji

2017-07-26

Animals perform near-optimal probabilistic inference in a wide range of psychophysical tasks. Probabilistic inference requires trial-to-trial representation of the uncertainties associated with task variables and subsequent use of this representation. Previous work has implemented such computations using neural networks with hand-crafted and task-dependent operations. We show that generic neural networks trained with a simple error-based learning rule perform near-optimal probabilistic inference in nine common psychophysical tasks. In a probabilistic categorization task, error-based learning in a generic network simultaneously explains a monkey's learning curve and the evolution of qualitative aspects of its choice behavior. In all tasks, the number of neurons required for a given level of performance grows sublinearly with the input population size, a substantial improvement on previous implementations of probabilistic inference. The trained networks develop a novel sparsity-based probabilistic population code. Our results suggest that probabilistic inference emerges naturally in generic neural networks trained with error-based learning rules.Behavioural tasks often require probability distributions to be inferred about task specific variables. Here, the authors demonstrate that generic neural networks can be trained using a simple error-based learning rule to perform such probabilistic computations efficiently without any need for task specific operations.

Peer-Assisted Learning in the Athletic Training Clinical Setting

Science.gov (United States)

Henning, Jolene M; Weidner, Thomas G; Jones, James

2006-01-01

Context: Athletic training educators often anecdotally suggest that athletic training students enhance their learning by teaching their peers. However, peer-assisted learning (PAL) has not been examined within athletic training education in order to provide evidence for its current use or as a pedagogic tool. Objective: To describe the prevalence of PAL in athletic training clinical education and to identify students' perceptions of PAL. Design: Descriptive. Setting: “The Athletic Training Student Seminar” at the National Athletic Trainers' Association 2002 Annual Meeting and Clinical Symposia. Patients or Other Participants: A convenience sample of 138 entry-level male and female athletic training students. Main Outcome Measure(s): Students' perceptions regarding the prevalence and benefits of and preferences for PAL were measured using the Athletic Training Peer-Assisted Learning Assessment Survey. The Survey is a self-report tool with 4 items regarding the prevalence of PAL and 7 items regarding perceived benefits and preferences. Results: A total of 66% of participants practiced a moderate to large amount of their clinical skills with other athletic training students. Sixty percent of students reported feeling less anxious when performing clinical skills on patients in front of other athletic training students than in front of their clinical instructors. Chi-square analysis revealed that 91% of students enrolled in Commission on Accreditation of Allied Health Education Programs–accredited athletic training education programs learned a minimal to small amount of clinical skills from their peers compared with 65% of students in Joint Review Committee on Educational Programs in Athletic Training–candidacy schools (χ2 3 = 14.57, P < .01). Multiple analysis of variance revealed significant interactions between sex and academic level on several items regarding benefits and preferences. Conclusions: According to athletic training students, PAL is occurring in

Quantized, piecewise linear filter network

DEFF Research Database (Denmark)

Sørensen, John Aasted

1993-01-01

A quantization based piecewise linear filter network is defined. A method for the training of this network based on local approximation in the input space is devised. The training is carried out by repeatedly alternating between vector quantization of the training set into quantization classes...... and equalization of the quantization classes linear filter mean square training errors. The equalization of the mean square training errors is carried out by adapting the boundaries between neighbor quantization classes such that the differences in mean square training errors are reduced...

African Network Operators Group (AfNOG) Training Workshops and ...

International Development Research Centre (IDRC) Digital Library (Canada)

The African Network Operators Group (AfNOG) is a forum for technical cooperation and coordination between African network operators and engineers from the region's universities, research institutions and industry. This year, AfNOG's training workshops and meetings will be held in Rabat, Morocco, between 24 May and 6 ...

Effects of Cognitive Training on Resting-State Functional Connectivity of Default Mode, Salience, and Central Executive Networks.

Science.gov (United States)

Cao, Weifang; Cao, Xinyi; Hou, Changyue; Li, Ting; Cheng, Yan; Jiang, Lijuan; Luo, Cheng; Li, Chunbo; Yao, Dezhong

2016-01-01

Neuroimaging studies have documented that aging can disrupt certain higher cognitive systems such as the default mode network (DMN), the salience network and the central executive network (CEN). The effect of cognitive training on higher cognitive systems remains unclear. This study used a 1-year longitudinal design to explore the cognitive training effect on three higher cognitive networks in healthy older adults. The community-living healthy older adults were divided into two groups: the multi-domain cognitive training group (24 sessions of cognitive training over a 3-months period) and the wait-list control group. All subjects underwent cognitive measurements and resting-state functional magnetic resonance imaging scanning at baseline and at 1 year after the training ended. We examined training-related changes in functional connectivity (FC) within and between three networks. Compared with the baseline, we observed maintained or increased FC within all three networks after training. The scans after training also showed maintained anti-correlation of FC between the DMN and CEN compared to the baseline. These findings demonstrated that cognitive training maintained or improved the functional integration within networks and the coupling between the DMN and CEN in older adults. Our findings suggested that multi-domain cognitive training can mitigate the aging-related dysfunction of higher cognitive networks.

Neutron spectrometry with artificial neural networks

International Nuclear Information System (INIS)

Vega C, H.R.; Hernandez D, V.M.; Manzanares A, E.; Rodriguez, J.M.; Mercado S, G.A.; Iniguez de la Torre Bayo, M.P.; Barquero, R.; Arteaga A, T.

2005-01-01

An artificial neural network has been designed to obtain the neutron spectra from the Bonner spheres spectrometer's count rates. The neural network was trained using 129 neutron spectra. These include isotopic neutron sources; reference and operational spectra from accelerators and nuclear reactors, spectra from mathematical functions as well as few energy groups and monoenergetic spectra. The spectra were transformed from lethargy to energy distribution and were re-bin ned to 31 energy groups using the MCNP 4C code. Re-binned spectra and UTA4 response matrix were used to calculate the expected count rates in Bonner spheres spectrometer. These count rates were used as input and the respective spectrum was used as output during neural network training. After training the network was tested with the Bonner spheres count rates produced by a set of neutron spectra. This set contains data used during network training as well as data not used. Training and testing was carried out in the Mat lab program. To verify the network unfolding performance the original and unfolded spectra were compared using the χ 2 -test and the total fluence ratios. The use of Artificial Neural Networks to unfold neutron spectra in neutron spectrometry is an alternative procedure that overcomes the drawbacks associated in this ill-conditioned problem. (Author)

Neural network segmentation of magnetic resonance images

International Nuclear Information System (INIS)

Frederick, B.

1990-01-01

Neural networks are well adapted to the task of grouping input patterns into subsets which share some similarity. Moreover, once trained, they can generalize their classification rules to classify new data sets. Sets of pixel intensities from magnetic resonance (MR) images provide a natural input to a neural network; by varying imaging parameters, MR images can reflect various independent physical parameters of tissues in their pixel intensities. A neural net can then be trained to classify physically similar tissue types based on sets of pixel intensities resulting from different imaging studies on the same subject. This paper reports that a neural network classifier for image segmentation was implanted on a Sun 4/60, and was tested on the task of classifying tissues of canine head MR images. Four images of a transaxial slice with different imaging sequences were taken as input to the network (three spin-echo images and an inversion recovery image). The training set consisted of 691 representative samples of gray matter, white matter, cerebrospinal fluid, bone, and muscle preclassified by a neuroscientist. The network was trained using a fast backpropagation algorithm to derive the decision criteria to classify any location in the image by its pixel intensities, and the image was subsequently segmented by the classifier

Transfer of Training: Adding Insight through Social Network Analysis

Science.gov (United States)

Van den Bossche, Piet; Segers, Mien

2013-01-01

This article reviews studies which apply a social network perspective to examine transfer of training. The theory behind social networks focuses on the interpersonal mechanisms and social structures that exist among interacting units such as people within an organization. A premise of this perspective is that individual's behaviors and outcomes…

Edge union of networks on the same vertex set

International Nuclear Information System (INIS)

Loe, Chuan Wen; Jensen, Henrik Jeldtoft

2013-01-01

Random network generators such as Erdős–Rényi, Watts–Strogatz and Barabási–Albert models are used as models to study real-world networks. Let G 1 (V, E 1 ) and G 2 (V, E 2 ) be two such networks on the same vertex set V. This paper studies the degree distribution and clustering coefficient of the resultant networks, G(V, E 1 ∪E 2 ). (paper)

Edge union of networks on the same vertex set

Science.gov (United States)

Loe, Chuan Wen; Jeldtoft Jensen, Henrik

2013-06-01

Random network generators such as Erdős-Rényi, Watts-Strogatz and Barabási-Albert models are used as models to study real-world networks. Let G1(V, E1) and G2(V, E2) be two such networks on the same vertex set V. This paper studies the degree distribution and clustering coefficient of the resultant networks, G(V, E1∪E2).

Feature Set Evaluation for Offline Handwriting Recognition Systems: Application to the Recurrent Neural Network Model.

Science.gov (United States)

Chherawala, Youssouf; Roy, Partha Pratim; Cheriet, Mohamed

2016-12-01

The performance of handwriting recognition systems is dependent on the features extracted from the word image. A large body of features exists in the literature, but no method has yet been proposed to identify the most promising of these, other than a straightforward comparison based on the recognition rate. In this paper, we propose a framework for feature set evaluation based on a collaborative setting. We use a weighted vote combination of recurrent neural network (RNN) classifiers, each trained with a particular feature set. This combination is modeled in a probabilistic framework as a mixture model and two methods for weight estimation are described. The main contribution of this paper is to quantify the importance of feature sets through the combination weights, which reflect their strength and complementarity. We chose the RNN classifier because of its state-of-the-art performance. Also, we provide the first feature set benchmark for this classifier. We evaluated several feature sets on the IFN/ENIT and RIMES databases of Arabic and Latin script, respectively. The resulting combination model is competitive with state-of-the-art systems.

Training set optimization under population structure in genomic selection.

Science.gov (United States)

Isidro, Julio; Jannink, Jean-Luc; Akdemir, Deniz; Poland, Jesse; Heslot, Nicolas; Sorrells, Mark E

2015-01-01

Population structure must be evaluated before optimization of the training set population. Maximizing the phenotypic variance captured by the training set is important for optimal performance. The optimization of the training set (TRS) in genomic selection has received much interest in both animal and plant breeding, because it is critical to the accuracy of the prediction models. In this study, five different TRS sampling algorithms, stratified sampling, mean of the coefficient of determination (CDmean), mean of predictor error variance (PEVmean), stratified CDmean (StratCDmean) and random sampling, were evaluated for prediction accuracy in the presence of different levels of population structure. In the presence of population structure, the most phenotypic variation captured by a sampling method in the TRS is desirable. The wheat dataset showed mild population structure, and CDmean and stratified CDmean methods showed the highest accuracies for all the traits except for test weight and heading date. The rice dataset had strong population structure and the approach based on stratified sampling showed the highest accuracies for all traits. In general, CDmean minimized the relationship between genotypes in the TRS, maximizing the relationship between TRS and the test set. This makes it suitable as an optimization criterion for long-term selection. Our results indicated that the best selection criterion used to optimize the TRS seems to depend on the interaction of trait architecture and population structure.

Optimal Path Choice in Railway Passenger Travel Network Based on Residual Train Capacity

Directory of Open Access Journals (Sweden)

Fei Dou

2014-01-01

Full Text Available Passenger’s optimal path choice is one of the prominent research topics in the field of railway passenger transport organization. More and more different train types are available, increasing path choices from departure to destination for travelers are unstoppable. However, travelers cannot avoid being confused when they hope to choose a perfect travel plan based on various travel time and cost constraints before departure. In this study, railway passenger travel network is constructed based on train timetable. Both the generalized cost function we developed and the residual train capacity are considered to be the foundation of path searching procedure. The railway passenger travel network topology is analyzed based on residual train capacity. Considering the total travel time, the total travel cost, and the total number of passengers, we propose an optimal path searching algorithm based on residual train capacity in railway passenger travel network. Finally, the rationale of the railway passenger travel network and the optimal path generation algorithm are verified positively by case study.

Enumeration of minimal stoichiometric precursor sets in metabolic networks.

Science.gov (United States)

Andrade, Ricardo; Wannagat, Martin; Klein, Cecilia C; Acuña, Vicente; Marchetti-Spaccamela, Alberto; Milreu, Paulo V; Stougie, Leen; Sagot, Marie-France

2016-01-01

What an organism needs at least from its environment to produce a set of metabolites, e.g. target(s) of interest and/or biomass, has been called a minimal precursor set. Early approaches to enumerate all minimal precursor sets took into account only the topology of the metabolic network (topological precursor sets). Due to cycles and the stoichiometric values of the reactions, it is often not possible to produce the target(s) from a topological precursor set in the sense that there is no feasible flux. Although considering the stoichiometry makes the problem harder, it enables to obtain biologically reasonable precursor sets that we call stoichiometric. Recently a method to enumerate all minimal stoichiometric precursor sets was proposed in the literature. The relationship between topological and stoichiometric precursor sets had however not yet been studied. Such relationship between topological and stoichiometric precursor sets is highlighted. We also present two algorithms that enumerate all minimal stoichiometric precursor sets. The first one is of theoretical interest only and is based on the above mentioned relationship. The second approach solves a series of mixed integer linear programming problems. We compared the computed minimal precursor sets to experimentally obtained growth media of several Escherichia coli strains using genome-scale metabolic networks. The results show that the second approach efficiently enumerates minimal precursor sets taking stoichiometry into account, and allows for broad in silico studies of strains or species interactions that may help to understand e.g. pathotype and niche-specific metabolic capabilities. sasita is written in Java, uses cplex as LP solver and can be downloaded together with all networks and input files used in this paper at http://www.sasita.gforge.inria.fr.

Reward-based training of recurrent neural networks for cognitive and value-based tasks.

Science.gov (United States)

Song, H Francis; Yang, Guangyu R; Wang, Xiao-Jing

2017-01-13

Trained neural network models, which exhibit features of neural activity recorded from behaving animals, may provide insights into the circuit mechanisms of cognitive functions through systematic analysis of network activity and connectivity. However, in contrast to the graded error signals commonly used to train networks through supervised learning, animals learn from reward feedback on definite actions through reinforcement learning. Reward maximization is particularly relevant when optimal behavior depends on an animal's internal judgment of confidence or subjective preferences. Here, we implement reward-based training of recurrent neural networks in which a value network guides learning by using the activity of the decision network to predict future reward. We show that such models capture behavioral and electrophysiological findings from well-known experimental paradigms. Our work provides a unified framework for investigating diverse cognitive and value-based computations, and predicts a role for value representation that is essential for learning, but not executing, a task.

Routing trains through railway junctions: A new set-packing approach

DEFF Research Database (Denmark)

Lusby, Richard Martin; Larsen, Jesper; Ryan, David

2011-01-01

The problem of routing trains through railway junctions is an integral part of railway operations. Large junctions are highly interconnected networks of track where multiple railway lines merge, intersect, and split. The number of possible routings makes this a very complicated problem. We show h...

Improved Extension Neural Network and Its Applications

Directory of Open Access Journals (Sweden)

Yu Zhou

2014-01-01

Full Text Available Extension neural network (ENN is a new neural network that is a combination of extension theory and artificial neural network (ANN. The learning algorithm of ENN is based on supervised learning algorithm. One of important issues in the field of classification and recognition of ENN is how to achieve the best possible classifier with a small number of labeled training data. Training data selection is an effective approach to solve this issue. In this work, in order to improve the supervised learning performance and expand the engineering application range of ENN, we use a novel data selection method based on shadowed sets to refine the training data set of ENN. Firstly, we use clustering algorithm to label the data and induce shadowed sets. Then, in the framework of shadowed sets, the samples located around each cluster centers (core data and the borders between clusters (boundary data are selected as training data. Lastly, we use selected data to train ENN. Compared with traditional ENN, the proposed improved ENN (IENN has a better performance. Moreover, IENN is independent of the supervised learning algorithms and initial labeled data. Experimental results verify the effectiveness and applicability of our proposed work.

Interaction of multiple networks modulated by the working memory training based on real-time fMRI

Science.gov (United States)

Shen, Jiahui; Zhang, Gaoyan; Zhu, Chaozhe; Yao, Li; Zhao, Xiaojie

2015-03-01

Neuroimaging studies of working memory training have identified the alteration of brain activity as well as the regional interactions within the functional networks such as central executive network (CEN) and default mode network (DMN). However, how the interaction within and between these multiple networks is modulated by the training remains unclear. In this paper, we examined the interaction of three training-induced brain networks during working memory training based on real-time functional magnetic resonance imaging (rtfMRI). Thirty subjects assigned to the experimental and control group respectively participated in two times training separated by seven days. Three networks including silence network (SN), CEN and DMN were identified by the training data with the calculated function connections within each network. Structural equation modeling (SEM) approach was used to construct the directional connectivity patterns. The results showed that the causal influences from the percent signal changes of target ROI to the SN were positively changed in both two groups, as well as the causal influence from the SN to CEN was positively changed in experimental group but negatively changed in control group from the SN to DMN. Further correlation analysis of the changes in each network with the behavioral improvements showed that the changes in SN were stronger positively correlated with the behavioral improvement of letter memory task. These findings indicated that the SN was not only a switch between the target ROI and the other networks in the feedback training but also an essential factor to the behavioral improvement.

Neutron spectrometry using artificial neural networks

International Nuclear Information System (INIS)

Vega-Carrillo, Hector Rene; Martin Hernandez-Davila, Victor; Manzanares-Acuna, Eduardo; Mercado Sanchez, Gema A.; Pilar Iniguez de la Torre, Maria; Barquero, Raquel; Palacios, Francisco; Mendez Villafane, Roberto; Arteaga Arteaga, Tarcicio; Manuel Ortiz Rodriguez, Jose

2006-01-01

An artificial neural network has been designed to obtain neutron spectra from Bonner spheres spectrometer count rates. The neural network was trained using 129 neutron spectra. These include spectra from isotopic neutron sources; reference and operational spectra from accelerators and nuclear reactors, spectra based on mathematical functions as well as few energy groups and monoenergetic spectra. The spectra were transformed from lethargy to energy distribution and were re-binned to 31 energy groups using the MCNP 4C code. The re-binned spectra and the UTA4 response matrix were used to calculate the expected count rates in Bonner spheres spectrometer. These count rates were used as input and their respective spectra were used as output during the neural network training. After training, the network was tested with the Bonner spheres count rates produced by folding a set of neutron spectra with the response matrix. This set contains data used during network training as well as data not used. Training and testing was carried out using the Matlab ( R) program. To verify the network unfolding performance, the original and unfolded spectra were compared using the root mean square error. The use of artificial neural networks to unfold neutron spectra in neutron spectrometry is an alternative procedure that overcomes the drawbacks associated with this ill-conditioned problem

Effect of creatine supplementation and drop-set resistance training in untrained aging adults.

Science.gov (United States)

Johannsmeyer, Sarah; Candow, Darren G; Brahms, C Markus; Michel, Deborah; Zello, Gordon A

2016-10-01

To investigate the effects of creatine supplementation and drop-set resistance training in untrained aging adults. Participants were randomized to one of two groups: Creatine (CR: n=14, 7 females, 7 males; 58.0±3.0yrs, 0.1g/kg/day of creatine+0.1g/kg/day of maltodextrin) or Placebo (PLA: n=17, 7 females, 10 males; age: 57.6±5.0yrs, 0.2g/kg/day of maltodextrin) during 12weeks of drop-set resistance training (3days/week; 2 sets of leg press, chest press, hack squat and lat pull-down exercises performed to muscle fatigue at 80% baseline 1-repetition maximum [1-RM] immediately followed by repetitions to muscle fatigue at 30% baseline 1-RM). Prior to and following training and supplementation, assessments were made for body composition, muscle strength, muscle endurance, tasks of functionality, muscle protein catabolism and diet. Drop-set resistance training improved muscle mass, muscle strength, muscle endurance and tasks of functionality (pcreatine to drop-set resistance training significantly increased body mass (p=0.002) and muscle mass (p=0.007) compared to placebo. Males on creatine increased muscle strength (lat pull-down only) to a greater extent than females on creatine (p=0.005). Creatine enabled males to resistance train at a greater capacity over time compared to males on placebo (p=0.049) and females on creatine (p=0.012). Males on creatine (p=0.019) and females on placebo (p=0.014) decreased 3-MH compared to females on creatine. The addition of creatine to drop-set resistance training augments the gains in muscle mass from resistance training alone. Creatine is more effective in untrained aging males compared to untrained aging females. Copyright © 2016 Elsevier Inc. All rights reserved.

Education and Training, and Knowledge Networks for Capacity-Building in Nuclear Security

International Nuclear Information System (INIS)

Mrabit, Khammar

2014-01-01

Conclusions: • Capacity Building (CB) is critical for States to establish and maintain effective and sustainable nuclear security regime. • IAEA is a worldwide platform promoting international cooperation for CB in nuclear security involving more than 160 countries and over 20 Organizations and Initiatives. • IAEA Division of Nuclear Security is ready to continue supporting States in developing their CB through: – Comprehensive Training Programme: more than 80 training events annually – International Nuclear Security Training and Support Centre Network (NSSC) – Comprehensive Education Programme – International Nuclear Security Network (INSEN)

An Issue of Boundary Value for Velocity and Training Overhead Using Cooperative MIMO Technique in Wireless Sensor Network

Directory of Open Access Journals (Sweden)

M. R. Islam

2011-06-01

Full Text Available A boundary value of velocity of data gathering node (DGN and a critical value for training overhead beyond which the cooperative communication in wireless sensor network will not be feasible is proposed in this paper. Multiple Input Multiple Outputs (MIMO cooperative communication is taken as an application. The performance in terms of energy efficiency and delay for a combination of two transmitting and two receiving antennas is analyzed. The results show that a set of critical value of velocity and training overhead pair is present for the long haul communication from the sensors to the data gathering node. Later a graphical relation between boundary value of training overhead and velocity is simulated. A mathematical relation between velocity and training overhead is also developed. The effects of several parameters on training overhead and velocity are analyzed.

Cut set-based risk and reliability analysis for arbitrarily interconnected networks

Science.gov (United States)

Wyss, Gregory D.

2000-01-01

Method for computing all-terminal reliability for arbitrarily interconnected networks such as the United States public switched telephone network. The method includes an efficient search algorithm to generate minimal cut sets for nonhierarchical networks directly from the network connectivity diagram. Efficiency of the search algorithm stems in part from its basis on only link failures. The method also includes a novel quantification scheme that likewise reduces computational effort associated with assessing network reliability based on traditional risk importance measures. Vast reductions in computational effort are realized since combinatorial expansion and subsequent Boolean reduction steps are eliminated through analysis of network segmentations using a technique of assuming node failures to occur on only one side of a break in the network, and repeating the technique for all minimal cut sets generated with the search algorithm. The method functions equally well for planar and non-planar networks.

Engine cylinder pressure reconstruction using crank kinematics and recurrently-trained neural networks

Science.gov (United States)

Bennett, C.; Dunne, J. F.; Trimby, S.; Richardson, D.

2017-02-01

A recurrent non-linear autoregressive with exogenous input (NARX) neural network is proposed, and a suitable fully-recurrent training methodology is adapted and tuned, for reconstructing cylinder pressure in multi-cylinder IC engines using measured crank kinematics. This type of indirect sensing is important for cost effective closed-loop combustion control and for On-Board Diagnostics. The challenge addressed is to accurately predict cylinder pressure traces within the cycle under generalisation conditions: i.e. using data not previously seen by the network during training. This involves direct construction and calibration of a suitable inverse crank dynamic model, which owing to singular behaviour at top-dead-centre (TDC), has proved difficult via physical model construction, calibration, and inversion. The NARX architecture is specialised and adapted to cylinder pressure reconstruction, using a fully-recurrent training methodology which is needed because the alternatives are too slow and unreliable for practical network training on production engines. The fully-recurrent Robust Adaptive Gradient Descent (RAGD) algorithm, is tuned initially using synthesised crank kinematics, and then tested on real engine data to assess the reconstruction capability. Real data is obtained from a 1.125 l, 3-cylinder, in-line, direct injection spark ignition (DISI) engine involving synchronised measurements of crank kinematics and cylinder pressure across a range of steady-state speed and load conditions. The paper shows that a RAGD-trained NARX network using both crank velocity and crank acceleration as input information, provides fast and robust training. By using the optimum epoch identified during RAGD training, acceptably accurate cylinder pressures, and especially accurate location-of-peak-pressure, can be reconstructed robustly under generalisation conditions, making it the most practical NARX configuration and recurrent training methodology for use on production engines.

Neutron spectra unfolding in Bonner spheres spectrometry using neural networks

International Nuclear Information System (INIS)

Kardan, M.R.; Setayeshi, S.; Koohi-Fayegh, R.; Ghiassi-Nejad, M.

2003-01-01

The neural network method has been used for the unfolding of neutron spectra in neutron spectrometry by Bonner spheres. A back propagation algorithm was used for training of neural networks 4mm x 4 mm bare LiI(Eu) and in a polyethylene sphere set: 2, 3, 4, 5, 6, 7, 8, 10, 12, 18 inch diameter have been used for unfolding of neutron spectra. Neural networks were trained by 199 sets of neutron spectra, which were subdivided into 6, 8, 10, 12, 15 and 20 energy bins and for each of them an appropriate neural network was designed and trained. The validation was performed by the 21 sets of neutron spectra. A neural network with 10 energy bins which had a mean value of error of 6% for dose equivalent estimation of spectra in the validation set showed the best results. The obtained results show that neural networks can be applied as an effective method for unfolding neutron spectra especially when the main target is neutron dosimetry. (author)

Impedance-Based Harmonic Instability Assessment in Multiple Electric Trains and Traction Network Interaction System

DEFF Research Database (Denmark)

Tao, Haidong; Hu, Haitao; Wang, Xiongfei

2018-01-01

This paper presents an impedance-based method to systematically investigate the interaction between multi-train and traction networks, focusing on evaluating the harmonic instability problems. Firstly, the interaction mechanism of multi-train and the traction network is represented as a feedback ...

Does rational selection of training and test sets improve the outcome of QSAR modeling?

Science.gov (United States)

Martin, Todd M; Harten, Paul; Young, Douglas M; Muratov, Eugene N; Golbraikh, Alexander; Zhu, Hao; Tropsha, Alexander

2012-10-22

Prior to using a quantitative structure activity relationship (QSAR) model for external predictions, its predictive power should be established and validated. In the absence of a true external data set, the best way to validate the predictive ability of a model is to perform its statistical external validation. In statistical external validation, the overall data set is divided into training and test sets. Commonly, this splitting is performed using random division. Rational splitting methods can divide data sets into training and test sets in an intelligent fashion. The purpose of this study was to determine whether rational division methods lead to more predictive models compared to random division. A special data splitting procedure was used to facilitate the comparison between random and rational division methods. For each toxicity end point, the overall data set was divided into a modeling set (80% of the overall set) and an external evaluation set (20% of the overall set) using random division. The modeling set was then subdivided into a training set (80% of the modeling set) and a test set (20% of the modeling set) using rational division methods and by using random division. The Kennard-Stone, minimal test set dissimilarity, and sphere exclusion algorithms were used as the rational division methods. The hierarchical clustering, random forest, and k-nearest neighbor (kNN) methods were used to develop QSAR models based on the training sets. For kNN QSAR, multiple training and test sets were generated, and multiple QSAR models were built. The results of this study indicate that models based on rational division methods generate better statistical results for the test sets than models based on random division, but the predictive power of both types of models are comparable.

Neural-network-directed alignment of optical systems using the laser-beam spatial filter as an example

Science.gov (United States)

Decker, Arthur J.; Krasowski, Michael J.; Weiland, Kenneth E.

1993-01-01

This report describes an effort at NASA Lewis Research Center to use artificial neural networks to automate the alignment and control of optical measurement systems. Specifically, it addresses the use of commercially available neural network software and hardware to direct alignments of the common laser-beam-smoothing spatial filter. The report presents a general approach for designing alignment records and combining these into training sets to teach optical alignment functions to neural networks and discusses the use of these training sets to train several types of neural networks. Neural network configurations used include the adaptive resonance network, the back-propagation-trained network, and the counter-propagation network. This work shows that neural networks can be used to produce robust sequencers. These sequencers can learn by example to execute the step-by-step procedures of optical alignment and also can learn adaptively to correct for environmentally induced misalignment. The long-range objective is to use neural networks to automate the alignment and operation of optical measurement systems in remote, harsh, or dangerous aerospace environments. This work also shows that when neural networks are trained by a human operator, training sets should be recorded, training should be executed, and testing should be done in a manner that does not depend on intellectual judgments of the human operator.

Application of neural networks to quantitative spectrometry analysis

International Nuclear Information System (INIS)

Pilato, V.; Tola, F.; Martinez, J.M.; Huver, M.

1999-01-01

Accurate quantitative analysis of complex spectra (fission and activation products), relies upon experts' knowledge. In some cases several hours, even days of tedious calculations are needed. This is because current software is unable to solve deconvolution problems when several rays overlap. We have shown that such analysis can be correctly handled by a neural network, and the procedure can be automated with minimum laboratory measurements for networks training, as long as all the elements of the analysed solution figure in the training set and provided that adequate scaling of input data is performed. Once the network has been trained, analysis is carried out in a few seconds. On submitting to a test between several well-known laboratories, where unknown quantities of 57 Co, 58 Co, 85 Sr, 88 Y, 131 I, 139 Ce, 141 Ce present in a sample had to be determined, the results yielded by our network classed it amongst the best. The method is described, including experimental device and measures, training set designing, relevant input parameters definition, input data scaling and networks training. Main results are presented together with a statistical model allowing networks error prediction

Effects of cluster vs. traditional plyometric training sets on maximal-intensity exercise performance.

Science.gov (United States)

Asadi, Abbas; Ramírez-Campillo, Rodrigo

2016-01-01

The aim of this study was to compare the effects of 6-week cluster versus traditional plyometric training sets on jumping ability, sprint and agility performance. Thirteen college students were assigned to a cluster sets group (N=6) or traditional sets group (N=7). Both training groups completed the same training program. The traditional group completed five sets of 20 repetitions with 2min of rest between sets each session, while the cluster group completed five sets of 20 [2×10] repetitions with 30/90-s rest each session. Subjects were evaluated for countermovement jump (CMJ), standing long jump (SLJ), t test, 20-m and 40-m sprint test performance before and after the intervention. Both groups had similar improvements (Psets methods resulted in greater adaptations in sprint performance, while the cluster sets method resulted in greater jump and agility adaptations. Copyright © 2016 The Lithuanian University of Health Sciences. Production and hosting by Elsevier Urban & Partner Sp. z o.o. All rights reserved.

Internal measuring models in trained neural networks for parameter estimation from images

NARCIS (Netherlands)

Feng, Tian-Jin; Feng, T.J.; Houkes, Z.; Korsten, Maarten J.; Spreeuwers, Lieuwe Jan

1992-01-01

The internal representations of 'learned' knowledge in neural networks are still poorly understood, even for backpropagation networks. The paper discusses a possible interpretation of learned knowledge of a network trained for parameter estimation from images. The outputs of the hidden layer are the

The Effects of Long-term Abacus Training on Topological Properties of Brain Functional Networks.

Science.gov (United States)

Weng, Jian; Xie, Ye; Wang, Chunjie; Chen, Feiyan

2017-08-18

Previous studies in the field of abacus-based mental calculation (AMC) training have shown that this training has the potential to enhance a wide variety of cognitive abilities. It can also generate specific changes in brain structure and function. However, there is lack of studies investigating the impact of AMC training on the characteristics of brain networks. In this study, utilizing graph-based network analysis, we compared topological properties of brain functional networks between an AMC group and a matched control group. Relative to the control group, the AMC group exhibited higher nodal degrees in bilateral calcarine sulcus and increased local efficiency in bilateral superior occipital gyrus and right cuneus. The AMC group also showed higher nodal local efficiency in right fusiform gyrus, which was associated with better math ability. However, no relationship was significant in the control group. These findings provide evidence that long-term AMC training may improve information processing efficiency in visual-spatial related regions, which extend our understanding of training plasticity at the brain network level.

A Control Simulation Method of High-Speed Trains on Railway Network with Irregular Influence

International Nuclear Information System (INIS)

Yang Lixing; Li Xiang; Li Keping

2011-01-01

Based on the discrete time method, an effective movement control model is designed for a group of highspeed trains on a rail network. The purpose of the model is to investigate the specific traffic characteristics of high-speed trains under the interruption of stochastic irregular events. In the model, the high-speed rail traffic system is supposed to be equipped with the moving-block signalling system to guarantee maximum traversing capacity of the railway. To keep the safety of trains' movements, some operational strategies are proposed to control the movements of trains in the model, including traction operation, braking operation, and entering-station operation. The numerical simulations show that the designed model can well describe the movements of high-speed trains on the rail network. The research results can provide the useful information not only for investigating the propagation features of relevant delays under the irregular disturbance but also for rerouting and rescheduling trains on the rail network. (general)

Pap-smear Classification Using Efficient Second Order Neural Network Training Algorithms

DEFF Research Database (Denmark)

Ampazis, Nikolaos; Dounias, George; Jantzen, Jan

2004-01-01

In this paper we make use of two highly efficient second order neural network training algorithms, namely the LMAM (Levenberg-Marquardt with Adaptive Momentum) and OLMAM (Optimized Levenberg-Marquardt with Adaptive Momentum), for the construction of an efficient pap-smear test classifier. The alg......In this paper we make use of two highly efficient second order neural network training algorithms, namely the LMAM (Levenberg-Marquardt with Adaptive Momentum) and OLMAM (Optimized Levenberg-Marquardt with Adaptive Momentum), for the construction of an efficient pap-smear test classifier...

Design and Implementation of Behavior Recognition System Based on Convolutional Neural Network

Directory of Open Access Journals (Sweden)

Yu Bo

2017-01-01

Full Text Available We build a set of human behavior recognition system based on the convolution neural network constructed for the specific human behavior in public places. Firstly, video of human behavior data set will be segmented into images, then we process the images by the method of background subtraction to extract moving foreground characters of body. Secondly, the training data sets are trained into the designed convolution neural network, and the depth learning network is constructed by stochastic gradient descent. Finally, the various behaviors of samples are classified and identified with the obtained network model, and the recognition results are compared with the current mainstream methods. The result show that the convolution neural network can study human behavior model automatically and identify human’s behaviors without any manually annotated trainings.

Estimating the similarity of alternative Affymetrix probe sets using transcriptional networks

Science.gov (United States)

2013-01-01

Background The usefulness of the data from Affymetrix microarray analysis depends largely on the reliability of the files describing the correspondence between probe sets, genes and transcripts. Particularly, when a gene is targeted by several probe sets, these files should give information about the similarity of each alternative probe set pair. Transcriptional networks integrate the multiple correlations that exist between all probe sets and supply much more information than a simple correlation coefficient calculated for two series of signals. In this study, we used the PSAWN (Probe Set Assignment With Networks) programme we developed to investigate whether similarity of alternative probe sets resulted in some specific properties. Findings PSAWNpy delivered a full textual description of each probe set and information on the number and properties of secondary targets. PSAWNml calculated the similarity of each alternative probe set pair and allowed finding relationships between similarity and localisation of probes in common transcripts or exons. Similar alternative probe sets had very low negative correlation, high positive correlation and similar neighbourhood overlap. Using these properties, we devised a test that allowed grouping similar probe sets in a given network. By considering several networks, additional information concerning the similarity reproducibility was obtained, which allowed defining the actual similarity of alternative probe set pairs. In particular, we calculated the common localisation of probes in exons and in known transcripts and we showed that similarity was correctly correlated with them. The information collected on all pairs of alternative probe sets in the most popular 3’ IVT Affymetrix chips is available in tabular form at http://bns.crbm.cnrs.fr/download.html. Conclusions These processed data can be used to obtain a finer interpretation when comparing microarray data between biological conditions. They are particularly well

Estimating the similarity of alternative Affymetrix probe sets using transcriptional networks.

Science.gov (United States)

Bellis, Michel

2013-03-21

The usefulness of the data from Affymetrix microarray analysis depends largely on the reliability of the files describing the correspondence between probe sets, genes and transcripts. Particularly, when a gene is targeted by several probe sets, these files should give information about the similarity of each alternative probe set pair. Transcriptional networks integrate the multiple correlations that exist between all probe sets and supply much more information than a simple correlation coefficient calculated for two series of signals. In this study, we used the PSAWN (Probe Set Assignment With Networks) programme we developed to investigate whether similarity of alternative probe sets resulted in some specific properties. PSAWNpy delivered a full textual description of each probe set and information on the number and properties of secondary targets. PSAWNml calculated the similarity of each alternative probe set pair and allowed finding relationships between similarity and localisation of probes in common transcripts or exons. Similar alternative probe sets had very low negative correlation, high positive correlation and similar neighbourhood overlap. Using these properties, we devised a test that allowed grouping similar probe sets in a given network. By considering several networks, additional information concerning the similarity reproducibility was obtained, which allowed defining the actual similarity of alternative probe set pairs. In particular, we calculated the common localisation of probes in exons and in known transcripts and we showed that similarity was correctly correlated with them. The information collected on all pairs of alternative probe sets in the most popular 3' IVT Affymetrix chips is available in tabular form at http://bns.crbm.cnrs.fr/download.html. These processed data can be used to obtain a finer interpretation when comparing microarray data between biological conditions. They are particularly well adapted for searching 3' alternative

Permanent Set of Cross-Linking Networks: Comparison of Theory with Molecular Dynamics Simulations

DEFF Research Database (Denmark)

Rottach, Dana R.; Curro, John G.; Budzien, Joanne

2006-01-01

The permanent set of cross-linking networks is studied by molecular dynamics. The uniaxial stress for a bead-spring polymer network is investigated as a function of strain and cross-link density history, where cross-links are introduced in unstrained and strained networks. The permanent set...

Practical neural network recipies in C++

CERN Document Server

Masters

2014-01-01

This text serves as a cookbook for neural network solutions to practical problems using C++. It will enable those with moderate programming experience to select a neural network model appropriate to solving a particular problem, and to produce a working program implementing that network. The book provides guidance along the entire problem-solving path, including designing the training set, preprocessing variables, training and validating the network, and evaluating its performance. Though the book is not intended as a general course in neural networks, no background in neural works is assum

Optimizing Intermodal Train Schedules with a Design Balanced Network Design Model

DEFF Research Database (Denmark)

Pedersen, Michael Berliner; Crainic, Teodor Gabriel

We present a modeling approach for optimizing intermodal trains schedules based on an infrastructure divided into time-dependent train paths. The formulation can be generalized to a capacitated multi commodity network design model with additional design balance constraints. We present a Tabu Search...

Food hygiene training in small to medium-sized care settings.

Science.gov (United States)

Seaman, Phillip; Eves, Anita

2008-10-01

Adoption of safe food handling practices is essential to effectively manage food safety. This study explores the impact of basic or foundation level food hygiene training on the attitudes and intentions of food handlers in care settings, using questionnaires based on the Theory of Planned Behaviour. Interviews were also conducted with food handlers and their managers to ascertain beliefs about the efficacy of, perceived barriers to, and relevance of food hygiene training. Most food handlers had undertaken formal food hygiene training; however, many who had not yet received training were preparing food, including high risk foods. Appropriate pre-training support and on-going supervision appeared to be lacking, thus limiting the effectiveness of training. Findings showed Subjective Norm to be the most significant influence on food handlers' intention to perform safe food handling practices, irrespective of training status, emphasising the role of important others in determining desirable behaviours.

SuperNeurons: Dynamic GPU Memory Management for Training Deep Neural Networks

OpenAIRE

Wang, Linnan; Ye, Jinmian; Zhao, Yiyang; Wu, Wei; Li, Ang; Song, Shuaiwen Leon; Xu, Zenglin; Kraska, Tim

2018-01-01

Going deeper and wider in neural architectures improves the accuracy, while the limited GPU DRAM places an undesired restriction on the network design domain. Deep Learning (DL) practitioners either need change to less desired network architectures, or nontrivially dissect a network across multiGPUs. These distract DL practitioners from concentrating on their original machine learning tasks. We present SuperNeurons: a dynamic GPU memory scheduling runtime to enable the network training far be...

Improving the Robustness of Deep Neural Networks via Stability Training

OpenAIRE

Zheng, Stephan; Song, Yang; Leung, Thomas; Goodfellow, Ian

2016-01-01

In this paper we address the issue of output instability of deep neural networks: small perturbations in the visual input can significantly distort the feature embeddings and output of a neural network. Such instability affects many deep architectures with state-of-the-art performance on a wide range of computer vision tasks. We present a general stability training method to stabilize deep networks against small input distortions that result from various types of common image processing, such...

Annotating gene sets by mining large literature collections with protein networks.

Science.gov (United States)

Wang, Sheng; Ma, Jianzhu; Yu, Michael Ku; Zheng, Fan; Huang, Edward W; Han, Jiawei; Peng, Jian; Ideker, Trey

2018-01-01

Analysis of patient genomes and transcriptomes routinely recognizes new gene sets associated with human disease. Here we present an integrative natural language processing system which infers common functions for a gene set through automatic mining of the scientific literature with biological networks. This system links genes with associated literature phrases and combines these links with protein interactions in a single heterogeneous network. Multiscale functional annotations are inferred based on network distances between phrases and genes and then visualized as an ontology of biological concepts. To evaluate this system, we predict functions for gene sets representing known pathways and find that our approach achieves substantial improvement over the conventional text-mining baseline method. Moreover, our system discovers novel annotations for gene sets or pathways without previously known functions. Two case studies demonstrate how the system is used in discovery of new cancer-related pathways with ontological annotations.

Adaptive training of neural networks for control of autonomous mobile robots

NARCIS (Netherlands)

Steur, E.; Vromen, T.; Nijmeijer, H.; Fossen, T.I.; Nijmeijer, H.; Pettersen, K.Y.

2017-01-01

We present an adaptive training procedure for a spiking neural network, which is used for control of a mobile robot. Because of manufacturing tolerances, any hardware implementation of a spiking neural network has non-identical nodes, which limit the performance of the controller. The adaptive

Assessment of Random Assignment in Training and Test Sets using Generalized Cluster Analysis Technique

Directory of Open Access Journals (Sweden)

Sorana D. BOLBOACĂ

2011-06-01

Full Text Available Aim: The properness of random assignment of compounds in training and validation sets was assessed using the generalized cluster technique. Material and Method: A quantitative Structure-Activity Relationship model using Molecular Descriptors Family on Vertices was evaluated in terms of assignment of carboquinone derivatives in training and test sets during the leave-many-out analysis. Assignment of compounds was investigated using five variables: observed anticancer activity and four structure descriptors. Generalized cluster analysis with K-means algorithm was applied in order to investigate if the assignment of compounds was or not proper. The Euclidian distance and maximization of the initial distance using a cross-validation with a v-fold of 10 was applied. Results: All five variables included in analysis proved to have statistically significant contribution in identification of clusters. Three clusters were identified, each of them containing both carboquinone derivatives belonging to training as well as to test sets. The observed activity of carboquinone derivatives proved to be normal distributed on every. The presence of training and test sets in all clusters identified using generalized cluster analysis with K-means algorithm and the distribution of observed activity within clusters sustain a proper assignment of compounds in training and test set. Conclusion: Generalized cluster analysis using the K-means algorithm proved to be a valid method in assessment of random assignment of carboquinone derivatives in training and test sets.

Electronic collaboration in dermatology resident training through social networking.

Science.gov (United States)

Meeks, Natalie M; McGuire, April L; Carroll, Bryan T

2017-04-01

The use of online educational resources and professional social networking sites is increasing. The field of dermatology is currently under-utilizing online social networking as a means of professional collaboration and sharing of training materials. In this study, we sought to assess the current structure of and satisfaction with dermatology resident education and gauge interest for a professional social networking site for educational collaboration. Two surveys-one for residents and one for faculty-were electronically distributed via the American Society for Dermatologic Surgery and Association of Professors of Dermatology (APD) listserves. The surveys confirmed that there is interest among dermatology residents and faculty in a dermatology professional networking site with the goal to enhance educational collaboration.

Physicists set new record for network data transfer

CERN Multimedia

2007-01-01

"An international team of physicists, computer scientists, and network engineers joined forces to set new records for sustained data transfer between storage systems durint the SuperComputing 2006 (SC06) Bandwidth Challenge (BWC). (3 pages)

A Dynamic Linear Hashing Method for Redundancy Management in Train Ethernet Consist Network

Directory of Open Access Journals (Sweden)

Xiaobo Nie

2016-01-01

Full Text Available Massive transportation systems like trains are considered critical systems because they use the communication network to control essential subsystems on board. Critical system requires zero recovery time when a failure occurs in a communication network. The newly published IEC62439-3 defines the high-availability seamless redundancy protocol, which fulfills this requirement and ensures no frame loss in the presence of an error. This paper adopts these for train Ethernet consist network. The challenge is management of the circulating frames, capable of dealing with real-time processing requirements, fast switching times, high throughout, and deterministic behavior. The main contribution of this paper is the in-depth analysis it makes of network parameters imposed by the application of the protocols to train control and monitoring system (TCMS and the redundant circulating frames discarding method based on a dynamic linear hashing, using the fastest method in order to resolve all the issues that are dealt with.

Degradation analysis in the estimation of photometric redshifts from non-representative training sets

Science.gov (United States)

Rivera, J. D.; Moraes, B.; Merson, A. I.; Jouvel, S.; Abdalla, F. B.; Abdalla, M. C. B.

2018-04-01

We perform an analysis of photometric redshifts estimated by using a non-representative training sets in magnitude space. We use the ANNz2 and GPz algorithms to estimate the photometric redshift both in simulations as well as in real data from the Sloan Digital Sky Survey (DR12). We show that for the representative case, the results obtained by using both algorithms have the same quality, either using magnitudes or colours as input. In order to reduce the errors when estimating the redshifts with a non-representative training set, we perform the training in colour space. We estimate the quality of our results by using a mock catalogue which is split samples cuts in the r-band between 19.4 manage to define a photometric estimator which fits well the spectroscopic distribution of galaxies in the mock testing set, but with a larger scatter. To complete this work, we perform an analysis of the impact on the detection of clusters via density of galaxies in a field by using the photometric redshifts obtained with a non-representative training set.

Contemporary social network sites: Relevance in anesthesiology teaching, training, and research

OpenAIRE

Rudrashish Haldar; Ashutosh Kaushal; Sukhen Samanta; Paurush Ambesh; Shashi Srivastava; Prabhat K Singh

2016-01-01

Objective: The phenomenal popularity of social networking sites has been used globally by medical professionals to boost professional associations and scientific developments. They have tremendous potential to forge professional liaisons, generate employment,upgrading skills and publicizing scientific achievements. We highlight the role of social networking mediums in influencing teaching, training and research in anaesthesiology. Background: The growth of social networking sites have been pr...

Communication skills training in dementia care: a systematic review of effectiveness, training content, and didactic methods in different care settings.

Science.gov (United States)

Eggenberger, Eva; Heimerl, Katharina; Bennett, Michael I

2013-03-01

Caring for and caring about people with dementia require specific communication skills. Healthcare professionals and family caregivers usually receive little training to enable them to meet the communicative needs of people with dementia. This review identifies existent interventions to enhance communication in dementia care in various care settings. We searched MEDLINE, AMED, EMBASE, PsychINFO, CINAHL, The Cochrane Library, Gerolit, and Web of Science for scientific articles reporting interventions in both English and German. An intervention was defined as communication skills training by means of face-to-face interaction with the aim of improving basic communicative skills. Both professional and family caregivers were included. The effectiveness of such training was analyzed. Different types of training were defined. Didactic methods, training content, and additional organizational features were qualitatively examined. This review included 12 trials totaling 831 persons with dementia, 519 professional caregivers, and 162 family caregivers. Most studies were carried out in the USA, the UK, and Germany. Eight studies took place in nursing homes; four studies were located in a home-care setting. No studies could be found in an acute-care setting. We provide a list of basic communicative principles for good communication in dementia care. Didactic methods included lectures, hands-on training, group discussions, and role-play. This review shows that communication skills training in dementia care significantly improves the quality of life and wellbeing of people with dementia and increases positive interactions in various care settings. Communication skills training shows significant impact on professional and family caregivers' communication skills, competencies, and knowledge. Additional organizational features improve the sustainability of communication interventions.

Pap-smear Classification Using Efficient Second Order Neural Network Training Algorithms

DEFF Research Database (Denmark)

Ampazis, Nikolaos; Dounias, George; Jantzen, Jan

2004-01-01

In this paper we make use of two highly efficient second order neural network training algorithms, namely the LMAM (Levenberg-Marquardt with Adaptive Momentum) and OLMAM (Optimized Levenberg-Marquardt with Adaptive Momentum), for the construction of an efficient pap-smear test classifier. The alg......In this paper we make use of two highly efficient second order neural network training algorithms, namely the LMAM (Levenberg-Marquardt with Adaptive Momentum) and OLMAM (Optimized Levenberg-Marquardt with Adaptive Momentum), for the construction of an efficient pap-smear test classifier....... The algorithms are methodologically similar, and are based on iterations of the form employed in the Levenberg-Marquardt (LM) method for non-linear least squares problems with the inclusion of an additional adaptive momentum term arising from the formulation of the training task as a constrained optimization...

Computer interpretation of thallium SPECT studies based on neural network analysis

Science.gov (United States)

Wang, David C.; Karvelis, K. C.

1991-06-01

A class of artificial intelligence (Al) programs known as neural networks are well suited to pattern recognition. A neural network is trained rather than programmed to recognize patterns. This differs from "expert system" Al programs in that it is not following an extensive set of rules determined by the programmer, but rather bases its decision on a gestalt interpretation of the image. The "bullseye" images from cardiac stress thallium tests performed on 50 male patients, as well as several simulated images were used to train the network. The network was able to accurately classify all patients in the training set. The network was then tested against 50 unknown patients and was able to correctly categorize 77% of the areas of ischemia and 92% of the areas of infarction. While not yet matching the ability of a trained physician, the neural network shows great promise in this area and has potential application in other areas of medical imaging.

Computer interpretation of thallium SPECT studies based on neural network analysis

International Nuclear Information System (INIS)

Wang, D.C.; Karvelis, K.C.

1991-01-01

This paper reports that a class of artificial intelligence (AI) programs known as neural-networks are well suited to pattern recognition. A neural network is trained rather than programmed to recognize patterns. This differs from expert system AI programs in that it is not following an extensive set of rules determined by the programmer, but rather bases its decision on a gestalt interpretation of the image. The bullseye images from cardiac stress thallium tests performed on 50 male patients, as well as several simulated images were used to train the network. The network was able to accurately classify all patients in the training set. The network was then tested against 50 unknown patients and was able to correctly categorize 77% of the areas of ischemia and 92% of the areas of infarction. While not yet matching the ability of the trained physician, the neural network shows great promise in this area and has potential application in other areas of medical imaging

Planning Training Loads for the 400 M Hurdles in Three-Month Mesocycles using Artificial Neural Networks.

Science.gov (United States)

Przednowek, Krzysztof; Iskra, Janusz; Wiktorowicz, Krzysztof; Krzeszowski, Tomasz; Maszczyk, Adam

2017-12-01

This paper presents a novel approach to planning training loads in hurdling using artificial neural networks. The neural models performed the task of generating loads for athletes' training for the 400 meters hurdles. All the models were calculated based on the training data of 21 Polish National Team hurdlers, aged 22.25 ± 1.96, competing between 1989 and 2012. The analysis included 144 training plans that represented different stages in the annual training cycle. The main contribution of this paper is to develop neural models for planning training loads for the entire career of a typical hurdler. In the models, 29 variables were used, where four characterized the runner and 25 described the training process. Two artificial neural networks were used: a multi-layer perceptron and a network with radial basis functions. To assess the quality of the models, the leave-one-out cross-validation method was used in which the Normalized Root Mean Squared Error was calculated. The analysis shows that the method generating the smallest error was the radial basis function network with nine neurons in the hidden layer. Most of the calculated training loads demonstrated a non-linear relationship across the entire competitive period. The resulting model can be used as a tool to assist a coach in planning training loads during a selected training period.

Planning Training Loads for The 400 M Hurdles in Three-Month Mesocycles Using Artificial Neural Networks

Directory of Open Access Journals (Sweden)

Przednowek Krzysztof

2017-12-01

Full Text Available This paper presents a novel approach to planning training loads in hurdling using artificial neural networks. The neural models performed the task of generating loads for athletes’ training for the 400 meters hurdles. All the models were calculated based on the training data of 21 Polish National Team hurdlers, aged 22.25 ± 1.96, competing between 1989 and 2012. The analysis included 144 training plans that represented different stages in the annual training cycle. The main contribution of this paper is to develop neural models for planning training loads for the entire career of a typical hurdler. In the models, 29 variables were used, where four characterized the runner and 25 described the training process. Two artificial neural networks were used: a multi-layer perceptron and a network with radial basis functions. To assess the quality of the models, the leave-one-out cross-validation method was used in which the Normalized Root Mean Squared Error was calculated. The analysis shows that the method generating the smallest error was the radial basis function network with nine neurons in the hidden layer. Most of the calculated training loads demonstrated a non-linear relationship across the entire competitive period. The resulting model can be used as a tool to assist a coach in planning training loads during a selected training period.

ON OPERATION OF 740 M LONG FREIGHT TRAINS ON CZECH TEN-T RAILWAY NETWORK

Directory of Open Access Journals (Sweden)

Michal Drábek

2016-09-01

Full Text Available Regulation (EU No 1315/2013 defines actual scope of core and comprehensive TEN-T network, including both networks for railway freight transport. For the core network, possibility to operate 740 m long freight trains is required. The aim of this paper is to analyse availability of appropriate overtaking tracks for 740 m long freight trains. Due to ETCS braking curves and odometry, such trains, after ETCS implementation, will require 780-800 m long overtaking tracks. For practical reasons (e.g. bypass lines, whole Czech railway TEN-T network is analysed. The overtaking track, whose occupation means influence on scheduled traffic or threat to boarding passengers, are excluded. The data was collected from station schemes from Collection of Official Requisites for 2015/16 Timetable, issued by SŽDC, Czech state Infrastructure Manager. Most of appropriate tracks are over 800 m long, but their density in the network and in particular directions varies considerably. For freight traffic, gradient of the line is important, so in the resulting figure, there are marked significant peaks for particular lines as well. Czech TEN-T lines are further segmented on the basis of number of tracks and their traffic character. Then, specific issues on overtaking or crossing of 740 m long freight trains are discussed. As a conclusion, for long-term development of Czech TEN-T lines, targeted investment is recommended not only for passenger railway, but also for freight railway. An attractive capacity offer for railway undertakings, which can stimulate freight traffic on European Rail freight corridors, can be represented by network-bound periodic freight train paths with suitable long overtaking tracks outside bottlenecks. After the overtaking by passenger trains, a freight train should run without stop through large node station or a bottleneck area. Before the sections with high gradients, coupling of additional locomotives should be connected with the overtaking

Can surgical simulation be used to train detection and classification of neural networks?

Science.gov (United States)

Zisimopoulos, Odysseas; Flouty, Evangello; Stacey, Mark; Muscroft, Sam; Giataganas, Petros; Nehme, Jean; Chow, Andre; Stoyanov, Danail

2017-10-01

Computer-assisted interventions (CAI) aim to increase the effectiveness, precision and repeatability of procedures to improve surgical outcomes. The presence and motion of surgical tools is a key information input for CAI surgical phase recognition algorithms. Vision-based tool detection and recognition approaches are an attractive solution and can be designed to take advantage of the powerful deep learning paradigm that is rapidly advancing image recognition and classification. The challenge for such algorithms is the availability and quality of labelled data used for training. In this Letter, surgical simulation is used to train tool detection and segmentation based on deep convolutional neural networks and generative adversarial networks. The authors experiment with two network architectures for image segmentation in tool classes commonly encountered during cataract surgery. A commercially-available simulator is used to create a simulated cataract dataset for training models prior to performing transfer learning on real surgical data. To the best of authors' knowledge, this is the first attempt to train deep learning models for surgical instrument detection on simulated data while demonstrating promising results to generalise on real data. Results indicate that simulated data does have some potential for training advanced classification methods for CAI systems.

Transfer Learning with Convolutional Neural Networks for Classification of Abdominal Ultrasound Images.

Science.gov (United States)

Cheng, Phillip M; Malhi, Harshawn S

2017-04-01

The purpose of this study is to evaluate transfer learning with deep convolutional neural networks for the classification of abdominal ultrasound images. Grayscale images from 185 consecutive clinical abdominal ultrasound studies were categorized into 11 categories based on the text annotation specified by the technologist for the image. Cropped images were rescaled to 256 × 256 resolution and randomized, with 4094 images from 136 studies constituting the training set, and 1423 images from 49 studies constituting the test set. The fully connected layers of two convolutional neural networks based on CaffeNet and VGGNet, previously trained on the 2012 Large Scale Visual Recognition Challenge data set, were retrained on the training set. Weights in the convolutional layers of each network were frozen to serve as fixed feature extractors. Accuracy on the test set was evaluated for each network. A radiologist experienced in abdominal ultrasound also independently classified the images in the test set into the same 11 categories. The CaffeNet network classified 77.3% of the test set images accurately (1100/1423 images), with a top-2 accuracy of 90.4% (1287/1423 images). The larger VGGNet network classified 77.9% of the test set accurately (1109/1423 images), with a top-2 accuracy of VGGNet was 89.7% (1276/1423 images). The radiologist classified 71.7% of the test set images correctly (1020/1423 images). The differences in classification accuracies between both neural networks and the radiologist were statistically significant (p convolutional neural networks may be used to construct effective classifiers for abdominal ultrasound images.

Social Networking in School Psychology Training Programs: A Survey of Faculty and Graduate Students

Science.gov (United States)

Pham, Andy V.; Goforth, Anisa N.; Segool, Natasha; Burt, Isaac

2014-01-01

The increasing use of social networking sites has become an emerging focus in school psychology training, policy, and research. The purpose of the current study is to present data from a survey on social networking among faculty and graduate students in school psychology training programs. A total of 110 faculty and 112 graduate students in school…

Training and Validating a Deep Convolutional Neural Network for Computer-Aided Detection and Classification of Abnormalities on Frontal Chest Radiographs.

Science.gov (United States)

Cicero, Mark; Bilbily, Alexander; Colak, Errol; Dowdell, Tim; Gray, Bruce; Perampaladas, Kuhan; Barfett, Joseph

2017-05-01

Convolutional neural networks (CNNs) are a subtype of artificial neural network that have shown strong performance in computer vision tasks including image classification. To date, there has been limited application of CNNs to chest radiographs, the most frequently performed medical imaging study. We hypothesize CNNs can learn to classify frontal chest radiographs according to common findings from a sufficiently large data set. Our institution's research ethics board approved a single-center retrospective review of 35,038 adult posterior-anterior chest radiographs and final reports performed between 2005 and 2015 (56% men, average age of 56, patient type: 24% inpatient, 39% outpatient, 37% emergency department) with a waiver for informed consent. The GoogLeNet CNN was trained using 3 graphics processing units to automatically classify radiographs as normal (n = 11,702) or into 1 or more of cardiomegaly (n = 9240), consolidation (n = 6788), pleural effusion (n = 7786), pulmonary edema (n = 1286), or pneumothorax (n = 1299). The network's performance was evaluated using receiver operating curve analysis on a test set of 2443 radiographs with the criterion standard being board-certified radiologist interpretation. Using 256 × 256-pixel images as input, the network achieved an overall sensitivity and specificity of 91% with an area under the curve of 0.964 for classifying a study as normal (n = 1203). For the abnormal categories, the sensitivity, specificity, and area under the curve, respectively, were 91%, 91%, and 0.962 for pleural effusion (n = 782), 82%, 82%, and 0.868 for pulmonary edema (n = 356), 74%, 75%, and 0.850 for consolidation (n = 214), 81%, 80%, and 0.875 for cardiomegaly (n = 482), and 78%, 78%, and 0.861 for pneumothorax (n = 167). Current deep CNN architectures can be trained with modest-sized medical data sets to achieve clinically useful performance at detecting and excluding common pathology on chest radiographs.

Self-consistent determination of the spike-train power spectrum in a neural network with sparse connectivity

Directory of Open Access Journals (Sweden)

Benjamin eDummer

2014-09-01

Full Text Available A major source of random variability in cortical networks is the quasi-random arrival of presynaptic action potentials from many other cells. In network studies as well as in the study of the response properties of single cells embedded in a network, synaptic background input is often approximated by Poissonian spike trains. However, the output statistics of the cells is in most cases far from being Poisson. This is inconsistent with the assumption of similar spike-train statistics for pre- and postsynaptic cells in a recurrent network. Here we tackle this problem for the popular class of integrate-and-fire neurons and study a self-consistent statistics of input and output spectra of neural spike trains. Instead of actually using a large network, we use an iterative scheme, in which we simulate a single neuron over several generations. In each of these generations, the neuron is stimulated with surrogate stochastic input that has a similar statistics as the output of the previous generation. For the surrogate input, we employ two distinct approximations: (i a superposition of renewal spike trains with the same interspike interval density as observed in the previous generation and (ii a Gaussian current with a power spectrum proportional to that observed in the previous generation. For input parameters that correspond to balanced input in the network, both the renewal and the Gaussian iteration procedure converge quickly and yield comparable results for the self-consistent spike-train power spectrum. We compare our results to large-scale simulations of a random sparsely connected network of leaky integrate-and-fire neurons (Brunel, J. Comp. Neurosci. 2000 and show that in the asynchronous regime close to a state of balanced synaptic input from the network, our iterative schemes provide excellent approximations to the autocorrelation of spike trains in the recurrent network.

Impact of real-time fMRI working memory feedback training on the interactions between three core brain networks.

Science.gov (United States)

Zhang, Qiushi; Zhang, Gaoyan; Yao, Li; Zhao, Xiaojie

2015-01-01

Working memory (WM) refers to the temporary holding and manipulation of information during the performance of a range of cognitive tasks, and WM training is a promising method for improving an individual's cognitive functions. Our previous work demonstrated that WM performance can be improved through self-regulation of dorsal lateral prefrontal cortex (PFC) activation using real-time functional magnetic resonance imaging (rtfMRI), which enables individuals to control local brain activities volitionally according to the neurofeedback. Furthermore, research concerning large-scale brain networks has demonstrated that WM training requires the engagement of several networks, including the central executive network (CEN), the default mode network (DMN) and the salience network (SN), and functional connectivity within the CEN and DMN can be changed by WM training. Although a switching role of the SN between the CEN and DMN has been demonstrated, it remains unclear whether WM training can affect the interactions between the three networks and whether a similar mechanism also exists during the training process. In this study, we investigated the dynamic functional connectivity between the three networks during the rtfMRI feedback training using independent component analysis (ICA) and correlation analysis. The results indicated that functional connectivity within and between the three networks were significantly enhanced by feedback training, and most of the changes were associated with the insula and correlated with behavioral improvements. These findings suggest that the insula plays a critical role in the reorganization of functional connectivity among the three networks induced by rtfMRI training and in WM performance, thus providing new insights into the mechanisms of high-level functions and the clinical treatment of related functional impairments.

The NNSYSID Toolbox - A MATLAB Toolbox for System Identification with Neural Networks

DEFF Research Database (Denmark)

Nørgård, Peter Magnus; Ravn, Ole; Hansen, Lars Kai

1996-01-01

To assist the identification of nonlinear dynamic systems, a set of tools has been developed for the MATLAB(R) environment. The tools include a number of different model structures, highly effective training algorithms, functions for validating trained networks, and pruning algorithms for determi......To assist the identification of nonlinear dynamic systems, a set of tools has been developed for the MATLAB(R) environment. The tools include a number of different model structures, highly effective training algorithms, functions for validating trained networks, and pruning algorithms...

High-Throughput Classification of Radiographs Using Deep Convolutional Neural Networks.

Science.gov (United States)

Rajkomar, Alvin; Lingam, Sneha; Taylor, Andrew G; Blum, Michael; Mongan, John

2017-02-01

The study aimed to determine if computer vision techniques rooted in deep learning can use a small set of radiographs to perform clinically relevant image classification with high fidelity. One thousand eight hundred eighty-five chest radiographs on 909 patients obtained between January 2013 and July 2015 at our institution were retrieved and anonymized. The source images were manually annotated as frontal or lateral and randomly divided into training, validation, and test sets. Training and validation sets were augmented to over 150,000 images using standard image manipulations. We then pre-trained a series of deep convolutional networks based on the open-source GoogLeNet with various transformations of the open-source ImageNet (non-radiology) images. These trained networks were then fine-tuned using the original and augmented radiology images. The model with highest validation accuracy was applied to our institutional test set and a publicly available set. Accuracy was assessed by using the Youden Index to set a binary cutoff for frontal or lateral classification. This retrospective study was IRB approved prior to initiation. A network pre-trained on 1.2 million greyscale ImageNet images and fine-tuned on augmented radiographs was chosen. The binary classification method correctly classified 100 % (95 % CI 99.73-100 %) of both our test set and the publicly available images. Classification was rapid, at 38 images per second. A deep convolutional neural network created using non-radiological images, and an augmented set of radiographs is effective in highly accurate classification of chest radiograph view type and is a feasible, rapid method for high-throughput annotation.

Effect of exercise order on the number of repeats and training volume in the tri-set training method

Directory of Open Access Journals (Sweden)

Waynne Ferreira de Faria

2016-05-01

Full Text Available DOI: http://dx.doi.org/10.5007/1980-0037.2016v18n2p187   Although the tri-set system is widely adopted by athletes and experienced weight training practitioners aimed at optimizing the metabolic overload, there are still few works in literature on the effect of exercise order manipulation on this training system. Therefore, this work was aimed at investigating the effect of exercise order manipulation on the number of repeats and training volume using the tri-set system for lower limbs. This is a randomized cross-over design study. The experimental group consisted of 14 healthy men (23.53 ± 5.40 years; 24.51 ± 2.96 kg/m2. Subjects were submitted to two experimental sessions at different exercise order for lower limbs: Sequence A: squat on guided bar, leg press 45° and bilateral leg extension; sequence B: bilateral leg extension, leg press 45° and squat on guided bar. Three sets to volitional fatigue in all exercises were performed, with intensity of 75% 1RM. Superiority for sequence B in the total number of repeats (70.14 ± 13 vs 60.93 ± 7.94 repeats, p = 0.004 and total training volume (9129.64 ± 2830.05 vs 8238.29 ± 2354.20 kg, p = 0.014 was observed. Based on the above, the performance of single-joint exercises before multi-joint exercises in the tri-set system adopted for lower limbs induced higher number of repeats and total training volume.

Picture this: Managed change and resistance in business network settings

DEFF Research Database (Denmark)

Kragh, Hanne; Andersen, Poul Houman

2009-01-01

This paper discusses change management in networks. The literature on business networks tends to downplay the role of managerial initiative in network change. The change management literature addresses such initiative, but with its single-firm perspective it overlooks the interdependence of network...... actors. In exploring the void between these two streams of literature, we deploy the concept of network pictures to discuss managed change in network settings. We analyze a change project from the furniture industry and address the consequences of attempting to manage change activities in a network...... context characterized by limited managerial authority over these activities. Our analysis suggests that change efforts unfold as a negotiated process during which the change project is re-negotiated to fit the multiple actor constituencies. The degree of overlap in the co-existing network pictures...

Neural network application to diesel generator diagnostics

International Nuclear Information System (INIS)

Logan, K.P.

1990-01-01

Diagnostic problems typically begin with the observation of some system behavior which is recognized as a deviation from the expected. The fundamental underlying process is one involving pattern matching cf observed symptoms to a set of compiled symptoms belonging to a fault-symptom mapping. Pattern recognition is often relied upon for initial fault detection and diagnosis. Parallel distributed processing (PDP) models employing neural network paradigms are known to be good pattern recognition devices. This paper describes the application of neural network processing techniques to the malfunction diagnosis of subsystems within a typical diesel generator configuration. Neural network models employing backpropagation learning were developed to correctly recognize fault conditions from the input diagnostic symptom patterns pertaining to various engine subsystems. The resulting network models proved to be excellent pattern recognizers for malfunction examples within the training set. The motivation for employing network models in lieu of a rule-based expert system, however, is related to the network's potential for generalizing malfunctions outside of the training set, as in the case of noisy or partial symptom patterns

Establishment of Oversea HRD Network and Operation of International Nuclear Education/Training Program

International Nuclear Information System (INIS)

Lee, E. J.; Min, B. J.; Han, K. W.

2008-02-01

The project deals with establishment of international network for human resources and the development of international nuclear education and training programs. The primary result is the establishment of KAERI International Nuclear R and D Academy as a new activity on cooperation for human resource development and building network. For this purpose, KAERI concluded the MOU with Vietnamese Universities and selected 3 students to provide Master and Ph. D. Courses in 2008. KAERI also held the 3rd World Nuclear University Summer Institute, in which some 150 international nuclear professionals attended for 6 weeks. Also, as part of regional networking, the Asian Network for Education in Nuclear Technology (ANENT) was promoted through development of a cyber platform and accomplishment the first IAEA e-training course. There were 3 kind of development activities for the international cooperation of human resources development. Firstly, the project provided training courses on nuclear energy development for the Egyptian Nuclear personnel under the bilateral cooperation. Secondly, the project published the English textbook and its lecture materials on introduction to nuclear engineering and fundamentals on OPR 1000 system technology. Lastly, the project developed a new KOICA training course on research reactor and radioisotope application technology to expand the KOICA sponsorship from 2008. The international nuclear education/training program had offered 15 courses to 314 people from 52 countries. In parallel, the project developed 11 kinds of lecturer materials and also developed 29 kinds of cyber lecturer materials. The operation of the International Nuclear Training and Education Center (INTEC) has contributed remarkably not only to the effective implementation of education/training activities of this project, but also to the promotion of other domestic and international activities of KAERI and other organizations

Establishment of Oversea HRD Network and Operation of International Nuclear Education/Training Program

Energy Technology Data Exchange (ETDEWEB)

Lee, E. J.; Min, B. J.; Han, K. W. (and others)

2008-02-15

The project deals with establishment of international network for human resources and the development of international nuclear education and training programs. The primary result is the establishment of KAERI International Nuclear R and D Academy as a new activity on cooperation for human resource development and building network. For this purpose, KAERI concluded the MOU with Vietnamese Universities and selected 3 students to provide Master and Ph. D. Courses in 2008. KAERI also held the 3rd World Nuclear University Summer Institute, in which some 150 international nuclear professionals attended for 6 weeks. Also, as part of regional networking, the Asian Network for Education in Nuclear Technology (ANENT) was promoted through development of a cyber platform and accomplishment the first IAEA e-training course. There were 3 kind of development activities for the international cooperation of human resources development. Firstly, the project provided training courses on nuclear energy development for the Egyptian Nuclear personnel under the bilateral cooperation. Secondly, the project published the English textbook and its lecture materials on introduction to nuclear engineering and fundamentals on OPR 1000 system technology. Lastly, the project developed a new KOICA training course on research reactor and radioisotope application technology to expand the KOICA sponsorship from 2008. The international nuclear education/training program had offered 15 courses to 314 people from 52 countries. In parallel, the project developed 11 kinds of lecturer materials and also developed 29 kinds of cyber lecturer materials. The operation of the International Nuclear Training and Education Center (INTEC) has contributed remarkably not only to the effective implementation of education/training activities of this project, but also to the promotion of other domestic and international activities of KAERI and other organizations.

Networked simulation for team training of Space Station astronauts, ground controllers, and scientists - A training and development environment

Science.gov (United States)

Hajare, Ankur R.; Wick, Daniel T.; Bovenzi, James J.

1991-01-01

The purpose of this paper is to describe plans for the Space Station Training Facility (SSTF) which has been designed to meet the envisioned training needs for Space Station Freedom. To meet these needs, the SSTF will integrate networked simulators with real-world systems in five training modes: Stand-Alone, Combined, Joint-Combined, Integrated, and Joint-Integrated. This paper describes the five training modes within the context of three training scenaries. In addition, this paper describes an authoring system which will support the rapid integration of new real-world system changes in the Space Station Freedom Program.

How large a training set is needed to develop a classifier for microarray data?

Science.gov (United States)

Dobbin, Kevin K; Zhao, Yingdong; Simon, Richard M

2008-01-01

A common goal of gene expression microarray studies is the development of a classifier that can be used to divide patients into groups with different prognoses, or with different expected responses to a therapy. These types of classifiers are developed on a training set, which is the set of samples used to train a classifier. The question of how many samples are needed in the training set to produce a good classifier from high-dimensional microarray data is challenging. We present a model-based approach to determining the sample size required to adequately train a classifier. It is shown that sample size can be determined from three quantities: standardized fold change, class prevalence, and number of genes or features on the arrays. Numerous examples and important experimental design issues are discussed. The method is adapted to address ex post facto determination of whether the size of a training set used to develop a classifier was adequate. An interactive web site for performing the sample size calculations is provided. We showed that sample size calculations for classifier development from high-dimensional microarray data are feasible, discussed numerous important considerations, and presented examples.

Interacting neural networks

Science.gov (United States)

Metzler, R.; Kinzel, W.; Kanter, I.

2000-08-01

Several scenarios of interacting neural networks which are trained either in an identical or in a competitive way are solved analytically. In the case of identical training each perceptron receives the output of its neighbor. The symmetry of the stationary state as well as the sensitivity to the used training algorithm are investigated. Two competitive perceptrons trained on mutually exclusive learning aims and a perceptron which is trained on the opposite of its own output are examined analytically. An ensemble of competitive perceptrons is used as decision-making algorithms in a model of a closed market (El Farol Bar problem or the Minority Game. In this game, a set of agents who have to make a binary decision is considered.); each network is trained on the history of minority decisions. This ensemble of perceptrons relaxes to a stationary state whose performance can be better than random.

Connected Dominating Set Based Topology Control in Wireless Sensor Networks

Science.gov (United States)

He, Jing

2012-01-01

Wireless Sensor Networks (WSNs) are now widely used for monitoring and controlling of systems where human intervention is not desirable or possible. Connected Dominating Sets (CDSs) based topology control in WSNs is one kind of hierarchical method to ensure sufficient coverage while reducing redundant connections in a relatively crowded network.…

EHV network operation, maintenance, organization and training

Energy Technology Data Exchange (ETDEWEB)

Gravier, J P [Electricite de France (EDF), 75 - Paris (France)

1994-12-31

The service interruptions of electricity have an ever increasing social and industrial impact, it is thus fundamental to operate the network to its best level of performances. To face these changing conditions, Electricite de France has consequently adapted its strategy to improve its organization for maintenance and operation, clarify the operation procedures and give further training to the staff. This work presents the above mentioned issues. (author) 2 figs.

Residents' perceived needs in communication skills training across in- and outpatient clinical settings.

Science.gov (United States)

Junod Perron, Noelle; Sommer, Johanna; Hudelson, Patricia; Demaurex, Florence; Luthy, Christophe; Louis-Simonet, Martine; Nendaz, Mathieu; De Grave, Willem; Dolmans, Diana; Van der Vleuten, Cees

2009-05-01

Residents' perceived needs in communication skills training are important to identify before designing context-specific training programmes, since learrners' perceived needs can influence the effectiveness of training. To explore residents' perceptions of their training needs and training experiences around communication skills, and whether these differ between residents training in inpatient and outpatient clinical settings. Four focus groups (FG) and a self-administered questionnaire were conducted with residents working in in- and outpatient medical service settings at a Swiss University Hospital. Focus groups explored residents' perceptions of their communication needs, their past training experiences and suggestions for future training programmes in communication skills. Transcripts were analysed in a thematic way using qualitative analytic approaches. All residents from both settings were asked to complete a questionnaire that queried their sociodemographics and amount of prior training in communication skills. In focus groups, outpatient residents felt that communication skills were especially useful in addressing chronic diseases and social issues. In contrast, inpatient residents emphasized the importance of good communication skills for dealing with family conflicts and end-of-life issues. Felt needs reflected residents' differing service priorities: outpatient residents saw the need for skills to structure the consultation and explore patients' perspectives in order to build therapeutic alliances, whereas inpatient residents wanted techniques to help them break bad news, provide information and increase their own well-being. The survey's overall response rate was 56%. Its data showed that outpatient residents received more training in communication skills and more of them than inpatient residents considered communication skills training to be useful (100% vs 74%). Outpatient residents' perceived needs in communication skills were more patient

Internal-state analysis in layered artificial neural network trained to categorize lung sounds

NARCIS (Netherlands)

Oud, M

2002-01-01

In regular use of artificial neural networks, only input and output states of the network are known to the user. Weight and bias values can be extracted but are difficult to interpret. We analyzed internal states of networks trained to map asthmatic lung sound spectra onto lung function parameters.

The Effect of Training Data Set Composition on the Performance of a Neural Image Caption Generator

Science.gov (United States)

2017-09-01

REPORT TYPE Technical Report 3. DATES COVERED (From - To) 4. TITLE AND SUBTITLE The Effect of Training Data Set Composition on the Performance of a...ARL-TR-8124 ● SEP 2017 US Army Research Laboratory The Effect of Training Data Set Composition on the Performance of a Neural...Laboratory The Effect of Training Data Set Composition on the Performance of a Neural Image Caption Generator by Abigail Wilson Montgomery Blair

Network-Based Coordination of Civil-Service Training: Lessons from the Case of Estonia

Directory of Open Access Journals (Sweden)

Metsma Merilin

2017-06-01

Full Text Available The focus of this article is on the coordination of civil-service training in a decentralized civil-service system. The Estonian case is studied. The article investigates network-based coordination, analyzes the power sources of the central coordinator and discusses the opportunities and limitations of creating coherence through network-type cooperation. The article concludes that the key power sources for the central coordinator are financial, human and technical resources paired with knowledge, leadership and commitment. The case study shows that, in a decentralized civil service system, a common understanding on training and development can be fostered by intense collaboration through networks.

Effects of training strategies implemented in a complex videogame on functional connectivity of attentional networks.

Science.gov (United States)

Voss, Michelle W; Prakash, Ruchika Shaurya; Erickson, Kirk I; Boot, Walter R; Basak, Chandramallika; Neider, Mark B; Simons, Daniel J; Fabiani, Monica; Gratton, Gabriele; Kramer, Arthur F

2012-01-02

We used the Space Fortress videogame, originally developed by cognitive psychologists to study skill acquisition, as a platform to examine learning-induced plasticity of interacting brain networks. Novice videogame players learned Space Fortress using one of two training strategies: (a) focus on all aspects of the game during learning (fixed priority), or (b) focus on improving separate game components in the context of the whole game (variable priority). Participants were scanned during game play using functional magnetic resonance imaging (fMRI), both before and after 20 h of training. As expected, variable priority training enhanced learning, particularly for individuals who initially performed poorly. Functional connectivity analysis revealed changes in brain network interaction reflective of more flexible skill learning and retrieval with variable priority training, compared to procedural learning and skill implementation with fixed priority training. These results provide the first evidence for differences in the interaction of large-scale brain networks when learning with different training strategies. Our approach and findings also provide a foundation for exploring the brain plasticity involved in transfer of trained abilities to novel real-world tasks such as driving, sport, or neurorehabilitation. Copyright © 2011 Elsevier Inc. All rights reserved.

A Web Based Educational Programming Logic Controller Training Set Based on Vocational High School Students' Demands

Directory of Open Access Journals (Sweden)

Abdullah Alper Efe

2018-01-01

Full Text Available The purpose of this study was to design and develop aProgramming Logic Controller Training Set according to vocational high school students’ educational needs. In this regard, by using the properties of distance education the proposed system supported “hands-on” PLC programming laboratory exercises in industrial automation area. The system allowed students to access and control the PLC training set remotely. For this purpose, researcher designed a web site to facilitate students’ interactivity and support PLC programming. In the training set, Induction Motor, Frequency Converter and Encoder tripart controlled by Siemens Simatic S7-200 PLC controller by the help of SIMATIC Step 7 Programming Software were used to make the system more effective and efficient. Moreover, training set included an IP camera system allowing to monitor devices and pilot application. By working with this novel remote accessible training set, students and researchers recieved a chance to inhere self paced learning experiences. Also, The PLC training set offered an effective learning enviroenment for distance education, which is based on presenting the content on the web and opening it to the online users and provided a safe and economical solution for multiple users in a workplace to enhance the quality of education with less overall cost.

Computing autocatalytic sets to unravel inconsistencies in metabolic network reconstructions

DEFF Research Database (Denmark)

Schmidt, R.; Waschina, S.; Boettger-Schmidt, D.

2015-01-01

, the method we report represents a powerful tool to identify inconsistencies in large-scale metabolic networks. AVAILABILITY AND IMPLEMENTATION: The method is available as source code on http://users.minet.uni-jena.de/ approximately m3kach/ASBIG/ASBIG.zip. CONTACT: christoph.kaleta@uni-jena.de SUPPLEMENTARY...... by inherent inconsistencies and gaps. RESULTS: Here we present a novel method to validate metabolic network reconstructions based on the concept of autocatalytic sets. Autocatalytic sets correspond to collections of metabolites that, besides enzymes and a growth medium, are required to produce all biomass...... components in a metabolic model. These autocatalytic sets are well-conserved across all domains of life, and their identification in specific genome-scale reconstructions allows us to draw conclusions about potential inconsistencies in these models. The method is capable of detecting inconsistencies, which...

The diagnosis of autism in community pediatric settings: does advanced training facilitate practice change?

Science.gov (United States)

Swanson, Amy R; Warren, Zachary E; Stone, Wendy L; Vehorn, Alison C; Dohrmann, Elizabeth; Humberd, Quentin

2014-07-01

The increased prevalence of autism spectrum disorder and documented benefits of early intensive intervention have created a need for flexible systems for determining eligibility for autism-specific services. This study evaluated the effectiveness of a training program designed to enhance autism spectrum disorder identification and assessment within community pediatric settings across the state. Twenty-seven pediatric providers participated in regional trainings across a 3.5-year period. Trainings provided clinicians with strategies for conducting relatively brief within-practice interactive assessments following positive autism spectrum disorder screenings. Program evaluation was measured approximately 1.5 years following training through (a) clinician self-reports of practice change and (b) blind diagnostic verification of a subset of children assessed. Pediatric providers participating in the training reported significant changes in screening and consultation practices following training, with a reported 85% increase in diagnostic identification of children with autism spectrum disorder within their own practice setting. In addition, substantial agreement (86%-93%) was found between pediatrician diagnostic judgments and independent, comprehensive blinded diagnostic evaluations. Collaborative training methods that allow autism spectrum disorder identification within broader community pediatric settings may help translate enhanced screening initiatives into more effective and efficient diagnosis and treatment. © The Author(s) 2013.

Neuromuscular and blood lactate responses to squat power training with different rest intervals between sets.

Science.gov (United States)

Martorelli, André; Bottaro, Martim; Vieira, Amilton; Rocha-Júnior, Valdinar; Cadore, Eduardo; Prestes, Jonato; Wagner, Dale; Martorelli, Saulo

2015-06-01

Studies investigating the effect of rest interval length (RI) between sets on neuromuscular performance and metabolic response during power training are scarce. Therefore, the purpose of this study was to compare maximal power output, muscular activity and blood lactate concentration following 1, 2 or 3 minutes RI between sets during a squat power training protocol. Twelve resistance-trained men (22.7 ± 3.2 years; 1.79 ± 0.08 cm; 81.8 ± 11.3 kg) performed 6 sets of 6 repetitions of squat exercise at 60% of their 1 repetition maximum. Peak and average power were obtained for each repetition and set using a linear position transducer. Muscular activity and blood lactate were measured pre and post-exercise session. There was no significant difference between RI on peak power and average power. However, peak power decreased 5.6%, 1.9%, and 5.9% after 6 sets using 1, 2 and 3 minutes of RI, respectively. Average power also decreased 10.5% (1 min), 2.6% (2 min), and 4.3% (3 min) after 6 sets. Blood lactate increased similarly during the three training sessions (1-min: 5.5 mMol, 2-min: 4.3 mMol, and 3-min: 4.0 mMol) and no significant changes were observed in the muscle activity after multiple sets, independent of RI length (pooled ES for 1-min: 0.47, 2-min: 0.65, and 3-min: 1.39). From a practical point of view, the results suggest that 1 to 2 minute of RI between sets during squat exercise may be sufficient to recover power output in a designed power training protocol. However, if training duration is malleable, we recommend 2 min of RI for optimal recovery and power output maintenance during the subsequent exercise sets. Key pointsThis study demonstrates that 1 minute of RI between sets is sufficient to maintain maximal power output during multiple sets of a power-based exercise when it is composed of few repetitions and the sets are not performed until failure. Therefore, a short RI should be considered when designing training programs for the development of

An Efficient Supervised Training Algorithm for Multilayer Spiking Neural Networks.

Science.gov (United States)

Xie, Xiurui; Qu, Hong; Liu, Guisong; Zhang, Malu; Kurths, Jürgen

2016-01-01

The spiking neural networks (SNNs) are the third generation of neural networks and perform remarkably well in cognitive tasks such as pattern recognition. The spike emitting and information processing mechanisms found in biological cognitive systems motivate the application of the hierarchical structure and temporal encoding mechanism in spiking neural networks, which have exhibited strong computational capability. However, the hierarchical structure and temporal encoding approach require neurons to process information serially in space and time respectively, which reduce the training efficiency significantly. For training the hierarchical SNNs, most existing methods are based on the traditional back-propagation algorithm, inheriting its drawbacks of the gradient diffusion and the sensitivity on parameters. To keep the powerful computation capability of the hierarchical structure and temporal encoding mechanism, but to overcome the low efficiency of the existing algorithms, a new training algorithm, the Normalized Spiking Error Back Propagation (NSEBP) is proposed in this paper. In the feedforward calculation, the output spike times are calculated by solving the quadratic function in the spike response model instead of detecting postsynaptic voltage states at all time points in traditional algorithms. Besides, in the feedback weight modification, the computational error is propagated to previous layers by the presynaptic spike jitter instead of the gradient decent rule, which realizes the layer-wised training. Furthermore, our algorithm investigates the mathematical relation between the weight variation and voltage error change, which makes the normalization in the weight modification applicable. Adopting these strategies, our algorithm outperforms the traditional SNN multi-layer algorithms in terms of learning efficiency and parameter sensitivity, that are also demonstrated by the comprehensive experimental results in this paper.

Training Feedforward Neural Networks Using Symbiotic Organisms Search Algorithm

Directory of Open Access Journals (Sweden)

Haizhou Wu

2016-01-01

Full Text Available Symbiotic organisms search (SOS is a new robust and powerful metaheuristic algorithm, which stimulates the symbiotic interaction strategies adopted by organisms to survive and propagate in the ecosystem. In the supervised learning area, it is a challenging task to present a satisfactory and efficient training algorithm for feedforward neural networks (FNNs. In this paper, SOS is employed as a new method for training FNNs. To investigate the performance of the aforementioned method, eight different datasets selected from the UCI machine learning repository are employed for experiment and the results are compared among seven metaheuristic algorithms. The results show that SOS performs better than other algorithms for training FNNs in terms of converging speed. It is also proven that an FNN trained by the method of SOS has better accuracy than most algorithms compared.

Application of Multiple-Population Genetic Algorithm in Optimizing the Train-Set Circulation Plan Problem

Directory of Open Access Journals (Sweden)

Yu Zhou

2017-01-01

Full Text Available The train-set circulation plan problem (TCPP belongs to the rolling stock scheduling (RSS problem and is similar to the aircraft routing problem (ARP in airline operations and the vehicle routing problem (VRP in the logistics field. However, TCPP involves additional complexity due to the maintenance constraint of train-sets: train-sets must conduct maintenance tasks after running for a certain time and distance. The TCPP is nondeterministic polynomial hard (NP-hard. There is no available algorithm that can obtain the optimal global solution, and many factors such as the utilization mode and the maintenance mode impact the solution of the TCPP. This paper proposes a train-set circulation optimization model to minimize the total connection time and maintenance costs and describes the design of an efficient multiple-population genetic algorithm (MPGA to solve this model. A realistic high-speed railway (HSR case is selected to verify our model and algorithm, and, then, a comparison of different algorithms is carried out. Furthermore, a new maintenance mode is proposed, and related implementation requirements are discussed.

Training the next generation of psychotraumatologists: COllaborative Network for Training and EXcellence in psychoTraumatology (CONTEXT)

Science.gov (United States)

Vallières, Frédérique; Hyland, Philip; Murphy, Jamie; Hansen, Maj; Shevlin, Mark; Elklit, Ask; Ceannt, Ruth; Armour, Cherie; Wiedemann, Nana; Munk, Mette; Dinesen, Cecilie; O’Hare, Geraldine; Cunningham, Twylla; Askerod, Ditte; Spitz, Pernille; Blackwell, Noeline; McCarthy, Angela; O’Dowd, Leonie; Scott, Shirley; Reid, Tracey; Mokake, Andreas; Halpin, Rory; Perera, Camila; Gleeson, Christina; Frost, Rachel; Flanagan, Natalie; Aldamman, Kinan; Tamrakar, Trina; Louison Vang, Maria; Sherwood, Larissa; Travers, Áine; Haahr-Pedersen, Ida; Walshe, Catherine; McDonagh, Tracey; Bramsen, Rikke Holm

2018-01-01

ABSTRACT In this paper we present a description of the Horizon2020, Marie Skłodowska-Curie Action funded, research and training programme CONTEXT: COllaborative Network for Training and EXcellence in psychoTraumatology. The three objectives of the programme are put forward, each of which refers to a key component of the CONTEXT programme. First, we summarize the 12 individual research projects that will take place across three priority populations: (i) refugees and asylum seekers, (ii) first responders, and (iii) perpetrators and survivors of childhood and gender-based violence. Second, we detail the mentoring and training programme central to CONTEXT. Finally, we describe how the research, together with the training, will contribute towards better policy, guidelines, and practice within the field of psychotraumatology. PMID:29372015

Training the next generation of psychotraumatologists: COllaborative Network for Training and EXcellence in psychoTraumatology (CONTEXT).

Science.gov (United States)

Vallières, Frédérique; Hyland, Philip; Murphy, Jamie; Hansen, Maj; Shevlin, Mark; Elklit, Ask; Ceannt, Ruth; Armour, Cherie; Wiedemann, Nana; Munk, Mette; Dinesen, Cecilie; O'Hare, Geraldine; Cunningham, Twylla; Askerod, Ditte; Spitz, Pernille; Blackwell, Noeline; McCarthy, Angela; O'Dowd, Leonie; Scott, Shirley; Reid, Tracey; Mokake, Andreas; Halpin, Rory; Perera, Camila; Gleeson, Christina; Frost, Rachel; Flanagan, Natalie; Aldamman, Kinan; Tamrakar, Trina; Louison Vang, Maria; Sherwood, Larissa; Travers, Áine; Haahr-Pedersen, Ida; Walshe, Catherine; McDonagh, Tracey; Bramsen, Rikke Holm

2018-01-01

In this paper we present a description of the Horizon2020, Marie Skłodowska-Curie Action funded, research and training programme CONTEXT: COllaborative Network for Training and EXcellence in psychoTraumatology. The three objectives of the programme are put forward, each of which refers to a key component of the CONTEXT programme. First, we summarize the 12 individual research projects that will take place across three priority populations: (i) refugees and asylum seekers, (ii) first responders, and (iii) perpetrators and survivors of childhood and gender-based violence. Second, we detail the mentoring and training programme central to CONTEXT. Finally, we describe how the research, together with the training, will contribute towards better policy, guidelines, and practice within the field of psychotraumatology.

Improving probe set selection for microbial community analysis by leveraging taxonomic information of training sequences

Directory of Open Access Journals (Sweden)

Jiang Tao

2011-10-01

Full Text Available Abstract Background Population levels of microbial phylotypes can be examined using a hybridization-based method that utilizes a small set of computationally-designed DNA probes targeted to a gene common to all. Our previous algorithm attempts to select a set of probes such that each training sequence manifests a unique theoretical hybridization pattern (a binary fingerprint to a probe set. It does so without taking into account similarity between training gene sequences or their putative taxonomic classifications, however. We present an improved algorithm for probe set selection that utilizes the available taxonomic information of training gene sequences and attempts to choose probes such that the resultant binary fingerprints cluster into real taxonomic groups. Results Gene sequences manifesting identical fingerprints with probes chosen by the new algorithm are more likely to be from the same taxonomic group than probes chosen by the previous algorithm. In cases where they are from different taxonomic groups, underlying DNA sequences of identical fingerprints are more similar to each other in probe sets made with the new versus the previous algorithm. Complete removal of large taxonomic groups from training data does not greatly decrease the ability of probe sets to distinguish those groups. Conclusions Probe sets made from the new algorithm create fingerprints that more reliably cluster into biologically meaningful groups. The method can readily distinguish microbial phylotypes that were excluded from the training sequences, suggesting novel microbes can also be detected.

Improving probe set selection for microbial community analysis by leveraging taxonomic information of training sequences.

Science.gov (United States)

Ruegger, Paul M; Della Vedova, Gianluca; Jiang, Tao; Borneman, James

2011-10-10

Population levels of microbial phylotypes can be examined using a hybridization-based method that utilizes a small set of computationally-designed DNA probes targeted to a gene common to all. Our previous algorithm attempts to select a set of probes such that each training sequence manifests a unique theoretical hybridization pattern (a binary fingerprint) to a probe set. It does so without taking into account similarity between training gene sequences or their putative taxonomic classifications, however. We present an improved algorithm for probe set selection that utilizes the available taxonomic information of training gene sequences and attempts to choose probes such that the resultant binary fingerprints cluster into real taxonomic groups. Gene sequences manifesting identical fingerprints with probes chosen by the new algorithm are more likely to be from the same taxonomic group than probes chosen by the previous algorithm. In cases where they are from different taxonomic groups, underlying DNA sequences of identical fingerprints are more similar to each other in probe sets made with the new versus the previous algorithm. Complete removal of large taxonomic groups from training data does not greatly decrease the ability of probe sets to distinguish those groups. Probe sets made from the new algorithm create fingerprints that more reliably cluster into biologically meaningful groups. The method can readily distinguish microbial phylotypes that were excluded from the training sequences, suggesting novel microbes can also be detected.

High-Speed Rail Train Timetabling Problem: A Time-Space Network Based Method with an Improved Branch-and-Price Algorithm

Directory of Open Access Journals (Sweden)

Bisheng He

2014-01-01

Full Text Available A time-space network based optimization method is designed for high-speed rail train timetabling problem to improve the service level of the high-speed rail. The general time-space path cost is presented which considers both the train travel time and the high-speed rail operation requirements: (1 service frequency requirement; (2 stopping plan adjustment; and (3 priority of train types. Train timetabling problem based on time-space path aims to minimize the total general time-space path cost of all trains. An improved branch-and-price algorithm is applied to solve the large scale integer programming problem. When dealing with the algorithm, a rapid branching and node selection for branch-and-price tree and a heuristic train time-space path generation for column generation are adopted to speed up the algorithm computation time. The computational results of a set of experiments on China’s high-speed rail system are presented with the discussions about the model validation, the effectiveness of the general time-space path cost, and the improved branch-and-price algorithm.

Performance of a visuomotor walking task in an augmented reality training setting

NARCIS (Netherlands)

Haarman, Juliet A.M.; Choi, Julia T.; Buurke, Jaap H.; Rietman, Johan S.; Reenalda, Jasper

2017-01-01

Visual cues can be used to train walking patterns. Here, we studied the performance and learning capacities of healthy subjects executing a high-precision visuomotor walking task, in an augmented reality training set-up. A beamer was used to project visual stepping targets on the walking surface of

Upper-Lower Bounds Candidate Sets Searching Algorithm for Bayesian Network Structure Learning

Directory of Open Access Journals (Sweden)

Guangyi Liu

2014-01-01

Full Text Available Bayesian network is an important theoretical model in artificial intelligence field and also a powerful tool for processing uncertainty issues. Considering the slow convergence speed of current Bayesian network structure learning algorithms, a fast hybrid learning method is proposed in this paper. We start with further analysis of information provided by low-order conditional independence testing, and then two methods are given for constructing graph model of network, which is theoretically proved to be upper and lower bounds of the structure space of target network, so that candidate sets are given as a result; after that a search and scoring algorithm is operated based on the candidate sets to find the final structure of the network. Simulation results show that the algorithm proposed in this paper is more efficient than similar algorithms with the same learning precision.

NGFATOS : national guidelines for first aid training in occupational settings

Science.gov (United States)

2002-05-01

NGFATOS is a course development guideline containing the essential elements of what can be considered safe, helpful and effective first aid training in occupational settings. This guide is intended for use by first aid program developers, institution...

Training-Image Based Geostatistical Inversion Using a Spatial Generative Adversarial Neural Network

Science.gov (United States)

Laloy, Eric; Hérault, Romain; Jacques, Diederik; Linde, Niklas

2018-01-01

Probabilistic inversion within a multiple-point statistics framework is often computationally prohibitive for high-dimensional problems. To partly address this, we introduce and evaluate a new training-image based inversion approach for complex geologic media. Our approach relies on a deep neural network of the generative adversarial network (GAN) type. After training using a training image (TI), our proposed spatial GAN (SGAN) can quickly generate 2-D and 3-D unconditional realizations. A key characteristic of our SGAN is that it defines a (very) low-dimensional parameterization, thereby allowing for efficient probabilistic inversion using state-of-the-art Markov chain Monte Carlo (MCMC) methods. In addition, available direct conditioning data can be incorporated within the inversion. Several 2-D and 3-D categorical TIs are first used to analyze the performance of our SGAN for unconditional geostatistical simulation. Training our deep network can take several hours. After training, realizations containing a few millions of pixels/voxels can be produced in a matter of seconds. This makes it especially useful for simulating many thousands of realizations (e.g., for MCMC inversion) as the relative cost of the training per realization diminishes with the considered number of realizations. Synthetic inversion case studies involving 2-D steady state flow and 3-D transient hydraulic tomography with and without direct conditioning data are used to illustrate the effectiveness of our proposed SGAN-based inversion. For the 2-D case, the inversion rapidly explores the posterior model distribution. For the 3-D case, the inversion recovers model realizations that fit the data close to the target level and visually resemble the true model well.

Maximizing the Lifetime of Wireless Sensor Networks Using Multiple Sets of Rendezvous

Directory of Open Access Journals (Sweden)

Bo Li

2015-01-01

Full Text Available In wireless sensor networks (WSNs, there is a “crowded center effect” where the energy of nodes located near a data sink drains much faster than other nodes resulting in a short network lifetime. To mitigate the “crowded center effect,” rendezvous points (RPs are used to gather data from other nodes. In order to prolong the lifetime of WSN further, we propose using multiple sets of RPs in turn to average the energy consumption of the RPs. The problem is how to select the multiple sets of RPs and how long to use each set of RPs. An optimal algorithm and a heuristic algorithm are proposed to address this problem. The optimal algorithm is highly complex and only suitable for small scale WSN. The performance of the proposed algorithms is evaluated through simulations. The simulation results indicate that the heuristic algorithm approaches the optimal one and that using multiple RP sets can significantly prolong network lifetime.

Statistical generation of training sets for measuring NO3(-), NH4(+) and major ions in natural waters using an ion selective electrode array.

Science.gov (United States)

Mueller, Amy V; Hemond, Harold F

2016-05-18

Knowledge of ionic concentrations in natural waters is essential to understand watershed processes. Inorganic nitrogen, in the form of nitrate and ammonium ions, is a key nutrient as well as a participant in redox, acid-base, and photochemical processes of natural waters, leading to spatiotemporal patterns of ion concentrations at scales as small as meters or hours. Current options for measurement in situ are costly, relying primarily on instruments adapted from laboratory methods (e.g., colorimetric, UV absorption); free-standing and inexpensive ISE sensors for NO3(-) and NH4(+) could be attractive alternatives if interferences from other constituents were overcome. Multi-sensor arrays, coupled with appropriate non-linear signal processing, offer promise in this capacity but have not yet successfully achieved signal separation for NO3(-) and NH4(+)in situ at naturally occurring levels in unprocessed water samples. A novel signal processor, underpinned by an appropriate sensor array, is proposed that overcomes previous limitations by explicitly integrating basic chemical constraints (e.g., charge balance). This work further presents a rationalized process for the development of such in situ instrumentation for NO3(-) and NH4(+), including a statistical-modeling strategy for instrument design, training/calibration, and validation. Statistical analysis reveals that historical concentrations of major ionic constituents in natural waters across New England strongly covary and are multi-modal. This informs the design of a statistically appropriate training set, suggesting that the strong covariance of constituents across environmental samples can be exploited through appropriate signal processing mechanisms to further improve estimates of minor constituents. Two artificial neural network architectures, one expanded to incorporate knowledge of basic chemical constraints, were tested to process outputs of a multi-sensor array, trained using datasets of varying degrees of

Bayesian model ensembling using meta-trained recurrent neural networks

NARCIS (Netherlands)

Ambrogioni, L.; Berezutskaya, Y.; Gü ç lü , U.; Borne, E.W.P. van den; Gü ç lü tü rk, Y.; Gerven, M.A.J. van; Maris, E.G.G.

2017-01-01

In this paper we demonstrate that a recurrent neural network meta-trained on an ensemble of arbitrary classification tasks can be used as an approximation of the Bayes optimal classifier. This result is obtained by relying on the framework of e-free approximate Bayesian inference, where the Bayesian

Detection of high-grade small bowel obstruction on conventional radiography with convolutional neural networks.

Science.gov (United States)

Cheng, Phillip M; Tejura, Tapas K; Tran, Khoa N; Whang, Gilbert

2018-05-01

The purpose of this pilot study is to determine whether a deep convolutional neural network can be trained with limited image data to detect high-grade small bowel obstruction patterns on supine abdominal radiographs. Grayscale images from 3663 clinical supine abdominal radiographs were categorized into obstructive and non-obstructive categories independently by three abdominal radiologists, and the majority classification was used as ground truth; 74 images were found to be consistent with small bowel obstruction. Images were rescaled and randomized, with 2210 images constituting the training set (39 with small bowel obstruction) and 1453 images constituting the test set (35 with small bowel obstruction). Weight parameters for the final classification layer of the Inception v3 convolutional neural network, previously trained on the 2014 Large Scale Visual Recognition Challenge dataset, were retrained on the training set. After training, the neural network achieved an AUC of 0.84 on the test set (95% CI 0.78-0.89). At the maximum Youden index (sensitivity + specificity-1), the sensitivity of the system for small bowel obstruction is 83.8%, with a specificity of 68.1%. The results demonstrate that transfer learning with convolutional neural networks, even with limited training data, may be used to train a detector for high-grade small bowel obstruction gas patterns on supine radiographs.

Computing single step operators of logic programming in radial basis function neural networks

Science.gov (United States)

Hamadneh, Nawaf; Sathasivam, Saratha; Choon, Ong Hong

2014-07-01

Logic programming is the process that leads from an original formulation of a computing problem to executable programs. A normal logic program consists of a finite set of clauses. A valuation I of logic programming is a mapping from ground atoms to false or true. The single step operator of any logic programming is defined as a function (Tp:I→I). Logic programming is well-suited to building the artificial intelligence systems. In this study, we established a new technique to compute the single step operators of logic programming in the radial basis function neural networks. To do that, we proposed a new technique to generate the training data sets of single step operators. The training data sets are used to build the neural networks. We used the recurrent radial basis function neural networks to get to the steady state (the fixed point of the operators). To improve the performance of the neural networks, we used the particle swarm optimization algorithm to train the networks.

Computing single step operators of logic programming in radial basis function neural networks

Energy Technology Data Exchange (ETDEWEB)

Hamadneh, Nawaf; Sathasivam, Saratha; Choon, Ong Hong [School of Mathematical Sciences, Universiti Sains Malaysia, 11800 USM, Penang (Malaysia)

2014-07-10

Logic programming is the process that leads from an original formulation of a computing problem to executable programs. A normal logic program consists of a finite set of clauses. A valuation I of logic programming is a mapping from ground atoms to false or true. The single step operator of any logic programming is defined as a function (T{sub p}:I→I). Logic programming is well-suited to building the artificial intelligence systems. In this study, we established a new technique to compute the single step operators of logic programming in the radial basis function neural networks. To do that, we proposed a new technique to generate the training data sets of single step operators. The training data sets are used to build the neural networks. We used the recurrent radial basis function neural networks to get to the steady state (the fixed point of the operators). To improve the performance of the neural networks, we used the particle swarm optimization algorithm to train the networks.

Computing single step operators of logic programming in radial basis function neural networks

International Nuclear Information System (INIS)

Hamadneh, Nawaf; Sathasivam, Saratha; Choon, Ong Hong

2014-01-01

Logic programming is the process that leads from an original formulation of a computing problem to executable programs. A normal logic program consists of a finite set of clauses. A valuation I of logic programming is a mapping from ground atoms to false or true. The single step operator of any logic programming is defined as a function (T p :I→I). Logic programming is well-suited to building the artificial intelligence systems. In this study, we established a new technique to compute the single step operators of logic programming in the radial basis function neural networks. To do that, we proposed a new technique to generate the training data sets of single step operators. The training data sets are used to build the neural networks. We used the recurrent radial basis function neural networks to get to the steady state (the fixed point of the operators). To improve the performance of the neural networks, we used the particle swarm optimization algorithm to train the networks

Mapping, Awareness, and Virtualization Network Administrator Training Tool (MAVNATT) Architecture and Framework

Science.gov (United States)

2015-06-01

unit may setup and teardown the entire tactical infrastructure multiple times per day. This tactical network administrator training is a critical...language and runs on Linux and Unix based systems. All provisioning is based around the Nagios Core application, a powerful backend solution for network...start up a large number of virtual machines quickly. CORE supports the simulation of fixed and mobile networks. CORE is open-source, written in Python

Training of reverse propagation neural networks applied to neutron dosimetry

International Nuclear Information System (INIS)

Hernandez P, C. F.; Martinez B, M. R.; Leon P, A. A.; Espinoza G, J. G.; Castaneda M, V. H.; Solis S, L. O.; Castaneda M, R.; Ortiz R, M.; Vega C, H. R.; Mendez V, R.; Gallego, E.; De Sousa L, M. A.

2016-10-01

Neutron dosimetry is of great importance in radiation protection as aims to provide dosimetric quantities to assess the magnitude of detrimental health effects due to exposure of neutron radiation. To quantify detriment to health is necessary to evaluate the dose received by the occupationally exposed personnel using different detection systems called dosimeters, which have very dependent responses to the energy distribution of neutrons. The neutron detection is a much more complex problem than the detection of charged particles, since it does not carry an electric charge, does not cause direct ionization and has a greater penetration power giving the possibility of interacting with matter in a different way. Because of this, various neutron detection systems have been developed, among which the Bonner spheres spectrometric system stands out due to the advantages that possesses, such as a wide range of energy, high sensitivity and easy operation. However, once obtained the counting rates, the problem lies in the neutron spectrum deconvolution, necessary for the calculation of the doses, using different mathematical methods such as Monte Carlo, maximum entropy, iterative methods among others, which present various difficulties that have motivated the development of new technologies. Nowadays, methods based on artificial intelligence technologies are being used to perform neutron dosimetry, mainly using the theory of artificial neural networks. In these new methods the need for spectrum reconstruction can be eliminated for the calculation of the doses. In this work an artificial neural network or reverse propagation was trained for the calculation of 15 equivalent doses from the counting rates of the Bonner spheres spectrometric system using a set of 7 spheres, one of 2 spheres and two of a single sphere of different sizes, testing different error values until finding the most appropriate. The optimum network topology was obtained through the robust design

Genetic Algorithm Optimized Neural Networks Ensemble as ...

African Journals Online (AJOL)

NJD

Improvements in neural network calibration models by a novel approach using neural network ensemble (NNE) for the simultaneous ... process by training a number of neural networks. .... Matlab® version 6.1 was employed for building principal component ... provide a fair simulation of calibration data set with some degree.

Reorganization of functional brain networks mediates the improvement of cognitive performance following real-time neurofeedback training of working memory.

Science.gov (United States)

Zhang, Gaoyan; Yao, Li; Shen, Jiahui; Yang, Yihong; Zhao, Xiaojie

2015-05-01

Working memory (WM) is essential for individuals' cognitive functions. Neuroimaging studies indicated that WM fundamentally relied on a frontoparietal working memory network (WMN) and a cinguloparietal default mode network (DMN). Behavioral training studies demonstrated that the two networks can be modulated by WM training. Different from the behavioral training, our recent study used a real-time functional MRI (rtfMRI)-based neurofeedback method to conduct WM training, demonstrating that WM performance can be significantly improved after successfully upregulating the activity of the target region of interest (ROI) in the left dorsolateral prefrontal cortex (Zhang et al., [2013]: PloS One 8:e73735); however, the neural substrate of rtfMRI-based WM training remains unclear. In this work, we assessed the intranetwork and internetwork connectivity changes of WMN and DMN during the training, and their correlations with the change of brain activity in the target ROI as well as with the improvement of post-training behavior. Our analysis revealed an "ROI-network-behavior" correlation relationship underlying the rtfMRI training. Further mediation analysis indicated that the reorganization of functional brain networks mediated the effect of self-regulation of the target brain activity on the improvement of cognitive performance following the neurofeedback training. The results of this study enhance our understanding of the neural basis of real-time neurofeedback and suggest a new direction to improve WM performance by regulating the functional connectivity in the WM related networks. © 2014 Wiley Periodicals, Inc.

Selected aspects of modelling of foreign exchange rates with neural networks

Directory of Open Access Journals (Sweden)

Václav Mastný

2005-01-01

Full Text Available This paper deals with forecasting of the high-frequency foreign exchange market with neural networks. The objective is to investigate some aspects of modelling with neural networks (impact of topology, size of training set and time horizon of the forecast on the performance of the network. The data used for the purpose of this paper contain 15-minute time series of US dollar against other major currencies, Japanese Yen, British Pound and Euro. The results show, that performance of the network in terms of correct directorial change is negatively influenced by increasing number of hidden neurons and decreasing size of training set. The performance of the network is influenced by sampling frequency.

Identifying a set of influential spreaders in complex networks

Science.gov (United States)

Zhang, Jian-Xiong; Chen, Duan-Bing; Dong, Qiang; Zhao, Zhi-Dan

2016-06-01

Identifying a set of influential spreaders in complex networks plays a crucial role in effective information spreading. A simple strategy is to choose top-r ranked nodes as spreaders according to influence ranking method such as PageRank, ClusterRank and k-shell decomposition. Besides, some heuristic methods such as hill-climbing, SPIN, degree discount and independent set based are also proposed. However, these approaches suffer from a possibility that some spreaders are so close together that they overlap sphere of influence or time consuming. In this report, we present a simply yet effectively iterative method named VoteRank to identify a set of decentralized spreaders with the best spreading ability. In this approach, all nodes vote in a spreader in each turn, and the voting ability of neighbors of elected spreader will be decreased in subsequent turn. Experimental results on four real networks show that under Susceptible-Infected-Recovered (SIR) and Susceptible-Infected (SI) models, VoteRank outperforms the traditional benchmark methods on both spreading rate and final affected scale. What’s more, VoteRank has superior computational efficiency.

Neutron spectrometry and dosimetry by means of Bonner spheres system and artificial neural networks applying robust design of artificial neural networks

Energy Technology Data Exchange (ETDEWEB)

Martinez B, M.R.; Ortiz R, J.M.; Vega C, H.R. [UAZ, Av. Ramon Lopez Velarde No. 801, 98000 Zacatecas (Mexico)

2006-07-01

An Artificial Neural Network has been designed, trained and tested to unfold neutron spectra and simultaneously to calculate equivalent doses. A set of 187 neutron spectra compiled by the International Atomic Energy Agency and 13 equivalent doses were used in the artificial neural network designed, trained and tested. In order to design the neural network was used the robust design of artificial neural networks methodology, which assures that the quality of the neural networks takes into account from the design stage. Unless previous works, here, for first time a group of neural networks were designed and trained to unfold 187 neutron spectra and at the same time to calculate 13 equivalent doses, starting from the count rates coming from the Bonner spheres system by using a systematic and experimental strategy. (Author)

Neutron spectrometry and dosimetry by means of Bonner spheres system and artificial neural networks applying robust design of artificial neural networks

International Nuclear Information System (INIS)

Martinez B, M.R.; Ortiz R, J.M.; Vega C, H.R.

2006-01-01

An Artificial Neural Network has been designed, trained and tested to unfold neutron spectra and simultaneously to calculate equivalent doses. A set of 187 neutron spectra compiled by the International Atomic Energy Agency and 13 equivalent doses were used in the artificial neural network designed, trained and tested. In order to design the neural network was used the robust design of artificial neural networks methodology, which assures that the quality of the neural networks takes into account from the design stage. Unless previous works, here, for first time a group of neural networks were designed and trained to unfold 187 neutron spectra and at the same time to calculate 13 equivalent doses, starting from the count rates coming from the Bonner spheres system by using a systematic and experimental strategy. (Author)

Increasing the effectiveness of instrumentation and control training programs using integrated training settings and a systematic approach to training

International Nuclear Information System (INIS)

McMahon, J.F.; Rakos, N.

1992-01-01

The performance of plant maintenance-related tasks assigned to instrumentation and control (I ampersand C) technicians can be broken down into physical skills required to do the task; resident knowledge of how to do the task; effect of maintenance on plant operating conditions; interactions with other plant organizations such as operations, radiation protection, and quality control; and knowledge of consequences of miss-action. A technician who has learned about the task in formal classroom presentations has not had the advantage of integrating that knowledge with the requisite physical and communication skills; hence, the first time these distinct and vital parts of the task equation are put together is on the job, during initial task performance. On-the-job training provides for the integration of skills and knowledge; however, this form of training is limited by plant conditions, availability of supporting players, and training experience levels of the personnel conducting the exercise. For licensed operations personnel, most nuclear utilities use formal classroom and a full-scope control room simulator to achieve the integration of skills and knowledge in a controlled training environment. TU Electric has taken that same approach into maintenance areas by including identical plant equipment in a laboratory setting for the large portion of training received by maintenance personnel at its Comanche Peak steam electric station. The policy of determining training needs and defining the scope of training by using the systematic approach to training has been highly effective and provided training at a reasonable cost (approximately $18.00/student contact hour)

Deformable image registration using convolutional neural networks

Science.gov (United States)

Eppenhof, Koen A. J.; Lafarge, Maxime W.; Moeskops, Pim; Veta, Mitko; Pluim, Josien P. W.

2018-03-01

Deformable image registration can be time-consuming and often needs extensive parameterization to perform well on a specific application. We present a step towards a registration framework based on a three-dimensional convolutional neural network. The network directly learns transformations between pairs of three-dimensional images. The outputs of the network are three maps for the x, y, and z components of a thin plate spline transformation grid. The network is trained on synthetic random transformations, which are applied to a small set of representative images for the desired application. Training therefore does not require manually annotated ground truth deformation information. The methodology is demonstrated on public data sets of inspiration-expiration lung CT image pairs, which come with annotated corresponding landmarks for evaluation of the registration accuracy. Advantages of this methodology are its fast registration times and its minimal parameterization.

Exercise order affects the total training volume and the ratings of perceived exertion in response to a super-set resistance training session

Directory of Open Access Journals (Sweden)

Balsamo S

2012-02-01

Full Text Available Sandor Balsamo1–3, Ramires Alsamir Tibana1,2,4, Dahan da Cunha Nascimento1,2, Gleyverton Landim de Farias1,2, Zeno Petruccelli1,2, Frederico dos Santos de Santana1,2, Otávio Vanni Martins1,2, Fernando de Aguiar1,2, Guilherme Borges Pereira4, Jéssica Cardoso de Souza4, Jonato Prestes41Department of Physical Education, Centro Universitário UNIEURO, Brasília, 2GEPEEFS (Resistance training and Health Research Group, Brasília/DF, 3Graduate Program in Medical Sciences, School of Medicine, Universidade de Brasília (UnB, Brasília, 4Graduation Program in Physical Education, Catholic University of Brasilia (UCB, Brasília/DF, BrazilAbstract: The super-set is a widely used resistance training method consisting of exercises for agonist and antagonist muscles with limited or no rest interval between them – for example, bench press followed by bent-over rows. In this sense, the aim of the present study was to compare the effects of different super-set exercise sequences on the total training volume. A secondary aim was to evaluate the ratings of perceived exertion and fatigue index in response to different exercise order. On separate testing days, twelve resistance-trained men, aged 23.0 ± 4.3 years, height 174.8 ± 6.75 cm, body mass 77.8 ± 13.27 kg, body fat 12.0% ± 4.7%, were submitted to a super-set method by using two different exercise orders: quadriceps (leg extension + hamstrings (leg curl (QH or hamstrings (leg curl + quadriceps (leg extension (HQ. Sessions consisted of three sets with a ten-repetition maximum load with 90 seconds rest between sets. Results revealed that the total training volume was higher for the HQ exercise order (P = 0.02 with lower perceived exertion than the inverse order (P = 0.04. These results suggest that HQ exercise order involving lower limbs may benefit practitioners interested in reaching a higher total training volume with lower ratings of perceived exertion compared with the leg extension plus leg curl

An Evolutionary Optimization Framework for Neural Networks and Neuromorphic Architectures

Energy Technology Data Exchange (ETDEWEB)

Schuman, Catherine D [ORNL; Plank, James [University of Tennessee (UT); Disney, Adam [University of Tennessee (UT); Reynolds, John [University of Tennessee (UT)

2016-01-01

As new neural network and neuromorphic architectures are being developed, new training methods that operate within the constraints of the new architectures are required. Evolutionary optimization (EO) is a convenient training method for new architectures. In this work, we review a spiking neural network architecture and a neuromorphic architecture, and we describe an EO training framework for these architectures. We present the results of this training framework on four classification data sets and compare those results to other neural network and neuromorphic implementations. We also discuss how this EO framework may be extended to other architectures.

Training set optimization and classifier performance in a top-down diabetic retinopathy screening system

Science.gov (United States)

Wigdahl, J.; Agurto, C.; Murray, V.; Barriga, S.; Soliz, P.

2013-03-01

Diabetic retinopathy (DR) affects more than 4.4 million Americans age 40 and over. Automatic screening for DR has shown to be an efficient and cost-effective way to lower the burden on the healthcare system, by triaging diabetic patients and ensuring timely care for those presenting with DR. Several supervised algorithms have been developed to detect pathologies related to DR, but little work has been done in determining the size of the training set that optimizes an algorithm's performance. In this paper we analyze the effect of the training sample size on the performance of a top-down DR screening algorithm for different types of statistical classifiers. Results are based on partial least squares (PLS), support vector machines (SVM), k-nearest neighbor (kNN), and Naïve Bayes classifiers. Our dataset consisted of digital retinal images collected from a total of 745 cases (595 controls, 150 with DR). We varied the number of normal controls in the training set, while keeping the number of DR samples constant, and repeated the procedure 10 times using randomized training sets to avoid bias. Results show increasing performance in terms of area under the ROC curve (AUC) when the number of DR subjects in the training set increased, with similar trends for each of the classifiers. Of these, PLS and k-NN had the highest average AUC. Lower standard deviation and a flattening of the AUC curve gives evidence that there is a limit to the learning ability of the classifiers and an optimal number of cases to train on.

Accident scenario diagnostics with neural networks

International Nuclear Information System (INIS)

Guo, Z.

1992-01-01

Nuclear power plants are very complex systems. The diagnoses of transients or accident conditions is very difficult because a large amount of information, which is often noisy, or intermittent, or even incomplete, need to be processed in real time. To demonstrate their potential application to nuclear power plants, neural networks axe used to monitor the accident scenarios simulated by the training simulator of TVA's Watts Bar Nuclear Power Plant. A self-organization network is used to compress original data to reduce the total number of training patterns. Different accident scenarios are closely related to different key parameters which distinguish one accident scenario from another. Therefore, the accident scenarios can be monitored by a set of small size neural networks, called modular networks, each one of which monitors only one assigned accident scenario, to obtain fast training and recall. Sensitivity analysis is applied to select proper input variables for modular networks

Designing Serious Games for getting transferable skills in training settings

Directory of Open Access Journals (Sweden)

Félix Buendía-García

2014-02-01

Full Text Available Nowadays, serious games are present in almost every educational context. The current work deals with the design of serious games oriented towards getting transferable skills in different kinds of training settings. These games can be a valuable way of engaging citizens and workers in the learning process by means of metaphors or similar mechanisms close to their user experience. They also contain an encouragement factor to uptake generic job competencies. An approach is proposed to develop this type of game by mixing traditional design steps with an instructional strategy to provide structured learning bites in training settings. Several game prototypes have been developed to test this approach in the context of courses for public employees. The obtained outcomes reveal the wider possibilities of serious games as educational resources, as well as the use of game achievements to evaluate the acquisition of transferable skills.

Fluid region segmentation in OCT images based on convolution neural network

Science.gov (United States)

Liu, Dong; Liu, Xiaoming; Fu, Tianyu; Yang, Zhou

2017-07-01

In the retinal image, characteristics of fluid have great significance for diagnosis in eye disease. In the clinical, the segmentation of fluid is usually conducted manually, but is time-consuming and the accuracy is highly depend on the expert's experience. In this paper, we proposed a segmentation method based on convolution neural network (CNN) for segmenting the fluid from fundus image. The B-scans of OCT are segmented into layers, and patches from specific region with annotation are used for training. After the data set being divided into training set and test set, network training is performed and a good segmentation result is obtained, which has a significant advantage over traditional methods such as threshold method.

Establishing Network Interaction between Resource Training Centers for People with Disabilities and Partner Universities

Directory of Open Access Journals (Sweden)

Panyukova S.V.,

2018-05-01

Full Text Available The paper focuses on the problem of accessibility and quality of higher education for students with disabilities. We describe our experience in organising network interaction between the MSUPE Resource and Training Center for Disabled People established in 2016-2017 and partner universities in ‘fixed territories’. The need for cooperation and network interaction arises from the high demand for the cooperation of efforts of leading experts, researchers, methodologists and instructors necessary for improving the quality and accessibility of higher education for persons with disabilities. The Resource and Training Center offers counseling for the partner universities, arranges advanced training for those responsible for teaching of the disabled, and offers specialized equipment for temporary use. In this article, we emphasize the importance of organizing network interactions with universities and social partners in order to ensure accessibility of higher education for students with disabilities.

The influence of negative training set size on machine learning-based virtual screening.

Science.gov (United States)

Kurczab, Rafał; Smusz, Sabina; Bojarski, Andrzej J

2014-01-01

The paper presents a thorough analysis of the influence of the number of negative training examples on the performance of machine learning methods. The impact of this rather neglected aspect of machine learning methods application was examined for sets containing a fixed number of positive and a varying number of negative examples randomly selected from the ZINC database. An increase in the ratio of positive to negative training instances was found to greatly influence most of the investigated evaluating parameters of ML methods in simulated virtual screening experiments. In a majority of cases, substantial increases in precision and MCC were observed in conjunction with some decreases in hit recall. The analysis of dynamics of those variations let us recommend an optimal composition of training data. The study was performed on several protein targets, 5 machine learning algorithms (SMO, Naïve Bayes, Ibk, J48 and Random Forest) and 2 types of molecular fingerprints (MACCS and CDK FP). The most effective classification was provided by the combination of CDK FP with SMO or Random Forest algorithms. The Naïve Bayes models appeared to be hardly sensitive to changes in the number of negative instances in the training set. In conclusion, the ratio of positive to negative training instances should be taken into account during the preparation of machine learning experiments, as it might significantly influence the performance of particular classifier. What is more, the optimization of negative training set size can be applied as a boosting-like approach in machine learning-based virtual screening.

Seismic signal auto-detecing from different features by using Convolutional Neural Network

Science.gov (United States)

Huang, Y.; Zhou, Y.; Yue, H.; Zhou, S.

2017-12-01

We try Convolutional Neural Network to detect some features of seismic data and compare their efficience. The features include whether a signal is seismic signal or noise and the arrival time of P and S phase and each feature correspond to a Convolutional Neural Network. We first use traditional STA/LTA to recongnize some events and then use templete matching to find more events as training set for the Neural Network. To make the training set more various, we add some noise to the seismic data and make some synthetic seismic data and noise. The 3-component raw signal and time-frequancy ananlyze are used as the input data for our neural network. Our Training is performed on GPUs to achieve efficient convergence. Our method improved the precision in comparison with STA/LTA and template matching. We will move to recurrent neural network to see if this kind network is better in detect P and S phase.

T-wave end detection using neural networks and Support Vector Machines.

Science.gov (United States)

Suárez-León, Alexander Alexeis; Varon, Carolina; Willems, Rik; Van Huffel, Sabine; Vázquez-Seisdedos, Carlos Román

2018-05-01

In this paper we propose a new approach for detecting the end of the T-wave in the electrocardiogram (ECG) using Neural Networks and Support Vector Machines. Both, Multilayer Perceptron (MLP) neural networks and Fixed-Size Least-Squares Support Vector Machines (FS-LSSVM) were used as regression algorithms to determine the end of the T-wave. Different strategies for selecting the training set such as random selection, k-means, robust clustering and maximum quadratic (Rényi) entropy were evaluated. Individual parameters were tuned for each method during training and the results are given for the evaluation set. A comparison between MLP and FS-LSSVM approaches was performed. Finally, a fair comparison of the FS-LSSVM method with other state-of-the-art algorithms for detecting the end of the T-wave was included. The experimental results show that FS-LSSVM approaches are more suitable as regression algorithms than MLP neural networks. Despite the small training sets used, the FS-LSSVM methods outperformed the state-of-the-art techniques. FS-LSSVM can be successfully used as a T-wave end detection algorithm in ECG even with small training set sizes. Copyright © 2018 Elsevier Ltd. All rights reserved.

Android malware detection based on evolutionary super-network

Science.gov (United States)

Yan, Haisheng; Peng, Lingling

2018-04-01

In the paper, an android malware detection method based on evolutionary super-network is proposed in order to improve the precision of android malware detection. Chi square statistics method is used for selecting characteristics on the basis of analyzing android authority. Boolean weighting is utilized for calculating characteristic weight. Processed characteristic vector is regarded as the system training set and test set; hyper edge alternative strategy is used for training super-network classification model, thereby classifying test set characteristic vectors, and it is compared with traditional classification algorithm. The results show that the detection method proposed in the paper is close to or better than traditional classification algorithm. The proposed method belongs to an effective Android malware detection means.

Train-Network Interactions and Stability Evaluation in High-Speed Railways--Part II: Influential Factors and Verifications

DEFF Research Database (Denmark)

Hu, Haitao; Tao, Haidong; Wang, Xiongfei

2018-01-01

Low-frequency oscillation (LFO), harmonic resonance and resonance instability phenomena happened in high speed railways (HSRs) are resulted from the interactions between multiple electric trains and traction network. A train-network interaction system and a unified impedance-based model......, catenary lines and autotransformers (ATs); 3) different numbers and positions of trains and railway lines will also be considered and discussed. In order to validate the theoretical results, the time-domain simulation and experiment system have been conducted. Finally, the differences and the relations...

Effects of cluster vs. traditional plyometric training sets on maximal-intensity exercise performance

Directory of Open Access Journals (Sweden)

Abbas Asadi

2016-01-01

Conclusions: Although both plyometric training methods improved lower body maximal-intensity exercise performance, the traditional sets methods resulted in greater adaptations in sprint performance, while the cluster sets method resulted in greater jump and agility adaptations.

Aging and Network Properties: Stability Over Time and Links with Learning during Working Memory Training

Directory of Open Access Journals (Sweden)

Alexandru D. Iordan

2018-01-01

Full Text Available Growing evidence suggests that healthy aging affects the configuration of large-scale functional brain networks. This includes reducing network modularity and local efficiency. However, the stability of these effects over time and their potential role in learning remain poorly understood. The goal of the present study was to further clarify previously reported age effects on “resting-state” networks, to test their reliability over time, and to assess their relation to subsequent learning during training. Resting-state fMRI data from 23 young (YA and 20 older adults (OA were acquired in 2 sessions 2 weeks apart. Graph-theoretic analyses identified both consistencies in network structure and differences in module composition between YA and OA, suggesting topological changes and less stability of functional network configuration with aging. Brain-wide, OA showed lower modularity and local efficiency compared to YA, consistent with the idea of age-related functional dedifferentiation, and these effects were replicable over time. At the level of individual networks, OA consistently showed greater participation and lower local efficiency and within-network connectivity in the cingulo-opercular network, as well as lower intra-network connectivity in the default-mode network and greater participation of the somato-sensorimotor network, suggesting age-related differential effects at the level of specialized brain modules. Finally, brain-wide network properties showed associations, albeit limited, with learning rates, as assessed with 10 days of computerized working memory training administered after the resting-state sessions, suggesting that baseline network configuration may influence subsequent learning outcomes. Identification of neural mechanisms associated with learning-induced plasticity is important for further clarifying whether and how such changes predict the magnitude and maintenance of training gains, as well as the extent and limits of

Development and Validation of a Deep Neural Network Model for Prediction of Postoperative In-hospital Mortality.

Science.gov (United States)

Lee, Christine K; Hofer, Ira; Gabel, Eilon; Baldi, Pierre; Cannesson, Maxime

2018-04-17

The authors tested the hypothesis that deep neural networks trained on intraoperative features can predict postoperative in-hospital mortality. The data used to train and validate the algorithm consists of 59,985 patients with 87 features extracted at the end of surgery. Feed-forward networks with a logistic output were trained using stochastic gradient descent with momentum. The deep neural networks were trained on 80% of the data, with 20% reserved for testing. The authors assessed improvement of the deep neural network by adding American Society of Anesthesiologists (ASA) Physical Status Classification and robustness of the deep neural network to a reduced feature set. The networks were then compared to ASA Physical Status, logistic regression, and other published clinical scores including the Surgical Apgar, Preoperative Score to Predict Postoperative Mortality, Risk Quantification Index, and the Risk Stratification Index. In-hospital mortality in the training and test sets were 0.81% and 0.73%. The deep neural network with a reduced feature set and ASA Physical Status classification had the highest area under the receiver operating characteristics curve, 0.91 (95% CI, 0.88 to 0.93). The highest logistic regression area under the curve was found with a reduced feature set and ASA Physical Status (0.90, 95% CI, 0.87 to 0.93). The Risk Stratification Index had the highest area under the receiver operating characteristics curve, at 0.97 (95% CI, 0.94 to 0.99). Deep neural networks can predict in-hospital mortality based on automatically extractable intraoperative data, but are not (yet) superior to existing methods.

Combining similarity in time and space for training set formation under concept drift

NARCIS (Netherlands)

Zliobaite, I.

2011-01-01

Concept drift is a challenge in supervised learning for sequential data. It describes a phenomenon when the data distributions change over time. In such a case accuracy of a classifier benefits from the selective sampling for training. We develop a method for training set selection, particularly

Modeling of steam generator in nuclear power plant using neural network ensemble

International Nuclear Information System (INIS)

Lee, S. K.; Lee, E. C.; Jang, J. W.

2003-01-01

Neural network is now being used in modeling the steam generator is known to be difficult due to the reverse dynamics. However, Neural network is prone to the problem of overfitting. This paper investigates the use of neural network combining methods to model steam generator water level and compares with single neural network. The results show that neural network ensemble is effective tool which can offer improved generalization, lower dependence of the training set and reduced training time

A multi-radio, multi-hop ad-hoc radio communication network for Communications-Based Train Control (CBTC): Introducing frequency separation for train-to-trackside communication

DEFF Research Database (Denmark)

Farooq, Jahanzeb; Bro, Lars; Karstensen, Rasmus Thystrup

2018-01-01

Communications-Based Train Control (CBTC) is a modern signalling system that uses radio communication to transfer train control information between train and wayside. The trackside networks in these systems are mostly based on conventionalinfrastructureWi-Fi(IEEE802.11).Itmeansatrain has to conti...

Holding-based network of nations based on listed energy companies: An empirical study on two-mode affiliation network of two sets of actors

Science.gov (United States)

Li, Huajiao; Fang, Wei; An, Haizhong; Gao, Xiangyun; Yan, Lili

2016-05-01

Economic networks in the real world are not homogeneous; therefore, it is important to study economic networks with heterogeneous nodes and edges to simulate a real network more precisely. In this paper, we present an empirical study of the one-mode derivative holding-based network constructed by the two-mode affiliation network of two sets of actors using the data of worldwide listed energy companies and their shareholders. First, we identify the primitive relationship in the two-mode affiliation network of the two sets of actors. Then, we present the method used to construct the derivative network based on the shareholding relationship between two sets of actors and the affiliation relationship between actors and events. After constructing the derivative network, we analyze different topological features on the node level, edge level and entire network level and explain the meanings of the different values of the topological features combining the empirical data. This study is helpful for expanding the usage of complex networks to heterogeneous economic networks. For empirical research on the worldwide listed energy stock market, this study is useful for discovering the inner relationships between the nations and regions from a new perspective.

Artificial Neural Networks and Instructional Technology.

Science.gov (United States)

Carlson, Patricia A.

1991-01-01

Artificial neural networks (ANN), part of artificial intelligence, are discussed. Such networks are fed sample cases (training sets), learn how to recognize patterns in the sample data, and use this experience in handling new cases. Two cognitive roles for ANNs (intelligent filters and spreading, associative memories) are examined. Prototypes…

Online Sequence Training of Recurrent Neural Networks with Connectionist Temporal Classification

OpenAIRE

Hwang, Kyuyeon; Sung, Wonyong

2015-01-01

Connectionist temporal classification (CTC) based supervised sequence training of recurrent neural networks (RNNs) has shown great success in many machine learning areas including end-to-end speech and handwritten character recognition. For the CTC training, however, it is required to unroll (or unfold) the RNN by the length of an input sequence. This unrolling requires a lot of memory and hinders a small footprint implementation of online learning or adaptation. Furthermore, the length of tr...

Identifying Learning Preferences in Vocational Education and Training Classroom Settings

Science.gov (United States)

Smith, Peter J.

2006-01-01

This research was designed to assess whether teachers and trainers of vocational learners noted and valued differences in individual learning preferences and, if so, how those differences were observed in natural classroom, workshop or other formal learning settings. Data were collected from six vocational education and training (VET) learning…

Routing Trains Through Railway Junctions: A New Set Packing Approach

DEFF Research Database (Denmark)

Lusby, Richard; Larsen, Jesper; Ryan, David

how the problem can be formulated as a set packing model. To exploit the structure of the problem we present a solution procedure which entails solving the dual of this formulation through the dynamic addition of violated cuts (primal variables). A discussion of the variable (train path) generation...

Bioinformatics Training Network (BTN): a community resource for bioinformatics trainers

DEFF Research Database (Denmark)

Schneider, Maria V.; Walter, Peter; Blatter, Marie-Claude

2012-01-01

and clearly tagged in relation to target audiences, learning objectives, etc. Ideally, they would also be peer reviewed, and easily and efficiently accessible for downloading. Here, we present the Bioinformatics Training Network (BTN), a new enterprise that has been initiated to address these needs and review...

Neural networks for sensor validation and plant monitoring

International Nuclear Information System (INIS)

Upadhyaya, B.R.; Eryurek, E.; Mathai, G.

1990-01-01

Sensor and process monitoring in power plants require the estimation of one or more process variables. Neural network paradigms are suitable for establishing general nonlinear relationships among a set of plant variables. Multiple-input multiple-output autoassociative networks can follow changes in plant-wide behavior. The backpropagation algorithm has been applied for training feedforward networks. A new and enhanced algorithm for training neural networks (BPN) has been developed and implemented in a VAX workstation. Operational data from the Experimental Breeder Reactor-II (EBR-II) have been used to study the performance of BPN. Several results of application to the EBR-II are presented

Effectiveness of behavioral skills training on staff performance in a job training setting for high-functioning adolescents with autism spectrum disorders

NARCIS (Netherlands)

Palmen, A.M.J.W.; Didden, H.C.M.; Korzilius, H.P.L.M.

2010-01-01

Few studies have focused on improving staff performance in naturalistic training settings for high-functioning adolescents with autism spectrum disorders. Behavioral skills training, consisting of group instruction and supervisory feedback, was used to improve staff performance on (a) providing

Integration of Neural Networks and Cellular Automata for Urban Planning

Institute of Scientific and Technical Information of China (English)

Anthony Gar-on Yeh; LI Xia

2004-01-01

This paper presents a new type of cellular automata (CA) model for the simulation of alternative land development using neural networks for urban planning. CA models can be regarded as a planning tool because they can generate alternative urban growth. Alternative development patterns can be formed by using different sets of parameter values in CA simulation. A critical issue is how to define parameter values for realistic and idealized simulation. This paper demonstrates that neural networks can simplify CA models but generate more plausible results. The simulation is based on a simple three-layer network with an output neuron to generate conversion probability. No transition rules are required for the simulation. Parameter values are automatically obtained from the training of network by using satellite remote sensing data. Original training data can be assessed and modified according to planning objectives. Alternative urban patterns can be easily formulated by using the modified training data sets rather than changing the model.

Neutron spectrum unfolding using neural networks

International Nuclear Information System (INIS)

Vega C, H.R.; Hernandez D, V.M.; Manzanares A, E.

2004-01-01

An artificial neural network has been designed to obtain the neutron spectra from the Bonner spheres spectrometer's count rates. The neural network was trained using a large set of neutron spectra compiled by the International Atomic Energy Agency. These include spectra from iso- topic neutron sources, reference and operational neutron spectra obtained from accelerators and nuclear reactors. The spectra were transformed from lethargy to energy distribution and were re-binned to 31 energy groups using the MCNP 4C code. Re-binned spectra and UTA4 matrix were used to calculate the expected count rates in Bonner spheres spectrometer. These count rates were used as input and correspondent spectrum was used as output during neural network training. The network has 7 input nodes, 56 neurons as hidden layer and 31 neurons in the output layer. After training the network was tested with the Bonner spheres count rates produced by twelve neutron spectra. The network allows unfolding the neutron spectrum from count rates measured with Bonner spheres. Good results are obtained when testing count rates belong to neutron spectra used during training, acceptable results are obtained for count rates obtained from actual neutron fields; however the network fails when count rates belong to monoenergetic neutron sources. (Author)

Deep Convolutional Neural Networks: Structure, Feature Extraction and Training

Directory of Open Access Journals (Sweden)

Namatēvs Ivars

2017-12-01

Full Text Available Deep convolutional neural networks (CNNs are aimed at processing data that have a known network like topology. They are widely used to recognise objects in images and diagnose patterns in time series data as well as in sensor data classification. The aim of the paper is to present theoretical and practical aspects of deep CNNs in terms of convolution operation, typical layers and basic methods to be used for training and learning. Some practical applications are included for signal and image classification. Finally, the present paper describes the proposed block structure of CNN for classifying crucial features from 3D sensor data.

Relationships between music training, speech processing, and word learning: a network perspective.

Science.gov (United States)

Elmer, Stefan; Jäncke, Lutz

2018-03-15

Numerous studies have documented the behavioral advantages conferred on professional musicians and children undergoing music training in processing speech sounds varying in the spectral and temporal dimensions. These beneficial effects have previously often been associated with local functional and structural changes in the auditory cortex (AC). However, this perspective is oversimplified, in that it does not take into account the intrinsic organization of the human brain, namely, neural networks and oscillatory dynamics. Therefore, we propose a new framework for extending these previous findings to a network perspective by integrating multimodal imaging, electrophysiology, and neural oscillations. In particular, we provide concrete examples of how functional and structural connectivity can be used to model simple neural circuits exerting a modulatory influence on AC activity. In addition, we describe how such a network approach can be used for better comprehending the beneficial effects of music training on more complex speech functions, such as word learning. © 2018 New York Academy of Sciences.

Practical guidelines for setting up neurosurgery skills training cadaver laboratory in India.

Science.gov (United States)

Suri, Ashish; Roy, Tara Sankar; Lalwani, Sanjeev; Deo, Rama Chandra; Tripathi, Manjul; Dhingra, Renu; Bhardwaj, Daya Nand; Sharma, Bhawani Shankar

2014-01-01

Though the necessity of cadaver dissection is felt by the medical fraternity, and described as early as 600 BC, in India, there are no practical guidelines available in the world literature for setting up a basic cadaver dissection laboratory for neurosurgery skills training. Hands-on dissection practice on microscopic and endoscopic procedures is essential in technologically demanding modern neurosurgery training where ethical issues, cost constraints, medico-legal pitfalls, and resident duty time restrictions have resulted in lesser opportunities to learn. Collaboration of anatomy, forensic medicine, and neurosurgery is essential for development of a workflow of cadaver procurement, preservation, storage, dissection, and disposal along with setting up the guidelines for ethical and legal concerns.

Training Deep Convolutional Neural Networks with Resistive Cross-Point Devices.

Science.gov (United States)

Gokmen, Tayfun; Onen, Murat; Haensch, Wilfried

2017-01-01

In a previous work we have detailed the requirements for obtaining maximal deep learning performance benefit by implementing fully connected deep neural networks (DNN) in the form of arrays of resistive devices. Here we extend the concept of Resistive Processing Unit (RPU) devices to convolutional neural networks (CNNs). We show how to map the convolutional layers to fully connected RPU arrays such that the parallelism of the hardware can be fully utilized in all three cycles of the backpropagation algorithm. We find that the noise and bound limitations imposed by the analog nature of the computations performed on the arrays significantly affect the training accuracy of the CNNs. Noise and bound management techniques are presented that mitigate these problems without introducing any additional complexity in the analog circuits and that can be addressed by the digital circuits. In addition, we discuss digitally programmable update management and device variability reduction techniques that can be used selectively for some of the layers in a CNN. We show that a combination of all those techniques enables a successful application of the RPU concept for training CNNs. The techniques discussed here are more general and can be applied beyond CNN architectures and therefore enables applicability of the RPU approach to a large class of neural network architectures.

Training Deep Convolutional Neural Networks with Resistive Cross-Point Devices

Science.gov (United States)

Gokmen, Tayfun; Onen, Murat; Haensch, Wilfried

2017-01-01

In a previous work we have detailed the requirements for obtaining maximal deep learning performance benefit by implementing fully connected deep neural networks (DNN) in the form of arrays of resistive devices. Here we extend the concept of Resistive Processing Unit (RPU) devices to convolutional neural networks (CNNs). We show how to map the convolutional layers to fully connected RPU arrays such that the parallelism of the hardware can be fully utilized in all three cycles of the backpropagation algorithm. We find that the noise and bound limitations imposed by the analog nature of the computations performed on the arrays significantly affect the training accuracy of the CNNs. Noise and bound management techniques are presented that mitigate these problems without introducing any additional complexity in the analog circuits and that can be addressed by the digital circuits. In addition, we discuss digitally programmable update management and device variability reduction techniques that can be used selectively for some of the layers in a CNN. We show that a combination of all those techniques enables a successful application of the RPU concept for training CNNs. The techniques discussed here are more general and can be applied beyond CNN architectures and therefore enables applicability of the RPU approach to a large class of neural network architectures. PMID:29066942

Prediction of residual stress for dissimilar metals welding at nuclear power plants using fuzzy neural network models

International Nuclear Information System (INIS)

Na, Man Gyun; Kim, Jin Weon; Lim, Dong Hyuk

2007-01-01

A fuzzy neural network model is presented to predict residual stress for dissimilar metal welding under various welding conditions. The fuzzy neural network model, which consists of a fuzzy inference system and a neuronal training system, is optimized by a hybrid learning method that combines a genetic algorithm to optimize the membership function parameters and a least squares method to solve the consequent parameters. The data of finite element analysis are divided into four data groups, which are split according to two end-section constraints and two prediction paths. Four fuzzy neural network models were therefore applied to the numerical data obtained from the finite element analysis for the two end-section constraints and the two prediction paths. The fuzzy neural network models were trained with the aid of a data set prepared for training (training data), optimized by means of an optimization data set and verified by means of a test data set that was different (independent) from the training data and the optimization data. The accuracy of fuzzy neural network models is known to be sufficiently accurate for use in an integrity evaluation by predicting the residual stress of dissimilar metal welding zones

A pre-trained convolutional neural network based method for thyroid nodule diagnosis.

Science.gov (United States)

Ma, Jinlian; Wu, Fa; Zhu, Jiang; Xu, Dong; Kong, Dexing

2017-01-01

In ultrasound images, most thyroid nodules are in heterogeneous appearances with various internal components and also have vague boundaries, so it is difficult for physicians to discriminate malignant thyroid nodules from benign ones. In this study, we propose a hybrid method for thyroid nodule diagnosis, which is a fusion of two pre-trained convolutional neural networks (CNNs) with different convolutional layers and fully-connected layers. Firstly, the two networks pre-trained with ImageNet database are separately trained. Secondly, we fuse feature maps learned by trained convolutional filters, pooling and normalization operations of the two CNNs. Finally, with the fused feature maps, a softmax classifier is used to diagnose thyroid nodules. The proposed method is validated on 15,000 ultrasound images collected from two local hospitals. Experiment results show that the proposed CNN based methods can accurately and effectively diagnose thyroid nodules. In addition, the fusion of the two CNN based models lead to significant performance improvement, with an accuracy of 83.02%±0.72%. These demonstrate the potential clinical applications of this method. Copyright © 2016 Elsevier B.V. All rights reserved.

Mindfulness Meditation Training and Executive Control Network Resting State Functional Connectivity: A Randomized Controlled Trial.

Science.gov (United States)

Taren, Adrienne A; Gianaros, Peter J; Greco, Carol M; Lindsay, Emily K; Fairgrieve, April; Brown, Kirk Warren; Rosen, Rhonda K; Ferris, Jennifer L; Julson, Erica; Marsland, Anna L; Creswell, J David

Mindfulness meditation training has been previously shown to enhance behavioral measures of executive control (e.g., attention, working memory, cognitive control), but the neural mechanisms underlying these improvements are largely unknown. Here, we test whether mindfulness training interventions foster executive control by strengthening functional connections between dorsolateral prefrontal cortex (dlPFC)-a hub of the executive control network-and frontoparietal regions that coordinate executive function. Thirty-five adults with elevated levels of psychological distress participated in a 3-day randomized controlled trial of intensive mindfulness meditation or relaxation training. Participants completed a resting state functional magnetic resonance imaging scan before and after the intervention. We tested whether mindfulness meditation training increased resting state functional connectivity (rsFC) between dlPFC and frontoparietal control network regions. Left dlPFC showed increased connectivity to the right inferior frontal gyrus (T = 3.74), right middle frontal gyrus (MFG) (T = 3.98), right supplementary eye field (T = 4.29), right parietal cortex (T = 4.44), and left middle temporal gyrus (T = 3.97, all p < .05) after mindfulness training relative to the relaxation control. Right dlPFC showed increased connectivity to right MFG (T = 4.97, p < .05). We report that mindfulness training increases rsFC between dlPFC and dorsal network (superior parietal lobule, supplementary eye field, MFG) and ventral network (right IFG, middle temporal/angular gyrus) regions. These findings extend previous work showing increased functional connectivity among brain regions associated with executive function during active meditation by identifying specific neural circuits in which rsFC is enhanced by a mindfulness intervention in individuals with high levels of psychological distress. Clinicaltrials.gov,NCT01628809.

An artificial neural network approach to reconstruct the source term of a nuclear accident

International Nuclear Information System (INIS)

Giles, J.; Palma, C. R.; Weller, P.

1997-01-01

This work makes use of one of the main features of artificial neural networks, which is their ability to 'learn' from sets of known input and output data. Indeed, a trained artificial neural network can be used to make predictions on the input data when the output is known, and this feedback process enables one to reconstruct the source term from field observations. With this aim, an artificial neural networks has been trained, using the projections of a segmented plume atmospheric dispersion model at fixed points, simulating a set of gamma detectors located outside the perimeter of a nuclear facility. The resulting set of artificial neural networks was used to determine the release fraction and rate for each of the noble gases, iodines and particulate fission products that could originate from a nuclear accident. Model projections were made using a large data set consisting of effective release height, release fraction of noble gases, iodines and particulate fission products, atmospheric stability, wind speed and wind direction. The model computed nuclide-specific gamma dose rates. The locations of the detectors were chosen taking into account both building shine and wake effects, and varied in distance between 800 and 1200 m from the reactor.The inputs to the artificial neural networks consisted of the measurements from the detector array, atmospheric stability, wind speed and wind direction; the outputs comprised a set of release fractions and heights. Once trained, the artificial neural networks was used to reconstruct the source term from the detector responses for data sets not used in training. The preliminary results are encouraging and show that the noble gases and particulate fission product release fractions are well determined

Design of cognitive engine for cognitive radio based on the rough sets and radial basis function neural network

Science.gov (United States)

Yang, Yanchao; Jiang, Hong; Liu, Congbin; Lan, Zhongli

2013-03-01

Cognitive radio (CR) is an intelligent wireless communication system which can dynamically adjust the parameters to improve system performance depending on the environmental change and quality of service. The core technology for CR is the design of cognitive engine, which introduces reasoning and learning methods in the field of artificial intelligence, to achieve the perception, adaptation and learning capability. Considering the dynamical wireless environment and demands, this paper proposes a design of cognitive engine based on the rough sets (RS) and radial basis function neural network (RBF_NN). The method uses experienced knowledge and environment information processed by RS module to train the RBF_NN, and then the learning model is used to reconfigure communication parameters to allocate resources rationally and improve system performance. After training learning model, the performance is evaluated according to two benchmark functions. The simulation results demonstrate the effectiveness of the model and the proposed cognitive engine can effectively achieve the goal of learning and reconfiguration in cognitive radio.

German MedicalTeachingNetwork (MDN) implementing national standards for teacher training.

Science.gov (United States)

Lammerding-Koeppel, M; Ebert, T; Goerlitz, A; Karsten, G; Nounla, C; Schmidt, S; Stosch, C; Dieter, P

2016-01-01

An increasing demand for proof of professionalism in higher education strives for quality assurance (QA) and improvement in medical education. A wide range of teacher trainings is available to medical staff in Germany. Cross-institutional approval of individual certificates is usually a difficult and time consuming task for institutions. In case of non-acceptance it may hinder medical teachers in their professional mobility. The faculties of medicine aimed to develop a comprehensive national framework, to promote standards for formal faculty development programmes across institutions and to foster professionalization of medical teaching. Addressing the above challenges in a joint approach, the faculties set up the national MedicalTeacherNetwork (MDN). Great importance is attributed to work out nationally concerted standards for faculty development and an agreed-upon quality control process across Germany. Medical teachers benefit from these advantages due to portability of faculty development credentials from one faculty of medicine to another within the MDN system. The report outlines the process of setting up the MDN and the national faculty development programme in Germany. Success factors, strengths and limitations are discussed from an institutional, individual and general perspective. Faculties engaged in similar developments might be encouraged to transfer the MDN concept to their countries.

Nonlinear Robust Observer-Based Fault Detection for Networked Suspension Control System of Maglev Train

Directory of Open Access Journals (Sweden)

Yun Li

2013-01-01

Full Text Available A fault detection approach based on nonlinear robust observer is designed for the networked suspension control system of Maglev train with random induced time delay. First, considering random bounded time-delay and external disturbance, the nonlinear model of the networked suspension control system is established. Then, a nonlinear robust observer is designed using the input of the suspension gap. And the estimate error is proved to be bounded with arbitrary precision by adopting an appropriate parameter. When sensor faults happen, the residual between the real states and the observer outputs indicates which kind of sensor failures occurs. Finally, simulation results using the actual parameters of CMS-04 maglev train indicate that the proposed method is effective for maglev train.

PEAK TRACKING WITH A NEURAL NETWORK FOR SPECTRAL RECOGNITION

NARCIS (Netherlands)

COENEGRACHT, PMJ; METTING, HJ; VANLOO, EM; SNOEIJER, GJ; DOORNBOS, DA

1993-01-01

A peak tracking method based on a simulated feed-forward neural network with back-propagation is presented. The network uses the normalized UV spectra and peak areas measured in one chromatogram for peak recognition. It suffices to train the network with only one set of spectra recorded in one

A deep level set method for image segmentation

OpenAIRE

Tang, Min; Valipour, Sepehr; Zhang, Zichen Vincent; Cobzas, Dana; MartinJagersand

2017-01-01

This paper proposes a novel image segmentation approachthat integrates fully convolutional networks (FCNs) with a level setmodel. Compared with a FCN, the integrated method can incorporatesmoothing and prior information to achieve an accurate segmentation.Furthermore, different than using the level set model as a post-processingtool, we integrate it into the training phase to fine-tune the FCN. Thisallows the use of unlabeled data during training in a semi-supervisedsetting. Using two types o...

Retrieving forest stand parameters from SAR backscatter data using a neural network trained by a canopy backscatter model

International Nuclear Information System (INIS)

Wang, Y.; Dong, D.

1997-01-01

It was possible to retrieve the stand mean dbh (tree trunk diameter at breast height) and stand density from the Jet Propulsion Laboratory (JPL) Airborne Synthetic Aperture Radar (AIRSAR) backscatter data by using threelayered perceptron neural networks (NNs). Two sets of NNs were trained by the Santa Barbara microwave canopy backscatter model. One set of the trained NNs was used to retrieve the stand mean dbh, and the other to retrieve the stand density. Each set of the NNs consisted of seven individual NNs for all possible combinations of one, two, and three radar wavelengths. Ground and multiple wavelength AIRSAR backscatter data from two ponderosa pine forest stands near Mt. Shasta, California (U.S.A.) were used to evaluate the accuracy of the retrievals. The r.m.s. and relative errors of the retrieval for stand mean dbh were 6.1 cm and 15.6 per cent for one stand (St2), and 3.1 cm and 6.7 per cent for the other stand (St11). The r.m.s. and relative errors of the retrieval for stand density were 71.2 treesha-1 and 23.0 per cent for St2, and 49.7 treesha-1 and 21.3 per cent for St11. (author)

Integrating Soft Set Theory and Fuzzy Linguistic Model to Evaluate the Performance of Training Simulation Systems.

Science.gov (United States)

Chang, Kuei-Hu; Chang, Yung-Chia; Chain, Kai; Chung, Hsiang-Yu

2016-01-01

The advancement of high technologies and the arrival of the information age have caused changes to the modern warfare. The military forces of many countries have replaced partially real training drills with training simulation systems to achieve combat readiness. However, considerable types of training simulation systems are used in military settings. In addition, differences in system set up time, functions, the environment, and the competency of system operators, as well as incomplete information have made it difficult to evaluate the performance of training simulation systems. To address the aforementioned problems, this study integrated analytic hierarchy process, soft set theory, and the fuzzy linguistic representation model to evaluate the performance of various training simulation systems. Furthermore, importance-performance analysis was adopted to examine the influence of saving costs and training safety of training simulation systems. The findings of this study are expected to facilitate applying military training simulation systems, avoiding wasting of resources (e.g., low utility and idle time), and providing data for subsequent applications and analysis. To verify the method proposed in this study, the numerical examples of the performance evaluation of training simulation systems were adopted and compared with the numerical results of an AHP and a novel AHP-based ranking technique. The results verified that not only could expert-provided questionnaire information be fully considered to lower the repetition rate of performance ranking, but a two-dimensional graph could also be used to help administrators allocate limited resources, thereby enhancing the investment benefits and training effectiveness of a training simulation system.

Exhaustively characterizing feasible logic models of a signaling network using Answer Set Programming.

Science.gov (United States)

Guziolowski, Carito; Videla, Santiago; Eduati, Federica; Thiele, Sven; Cokelaer, Thomas; Siegel, Anne; Saez-Rodriguez, Julio

2013-09-15

Logic modeling is a useful tool to study signal transduction across multiple pathways. Logic models can be generated by training a network containing the prior knowledge to phospho-proteomics data. The training can be performed using stochastic optimization procedures, but these are unable to guarantee a global optima or to report the complete family of feasible models. This, however, is essential to provide precise insight in the mechanisms underlaying signal transduction and generate reliable predictions. We propose the use of Answer Set Programming to explore exhaustively the space of feasible logic models. Toward this end, we have developed caspo, an open-source Python package that provides a powerful platform to learn and characterize logic models by leveraging the rich modeling language and solving technologies of Answer Set Programming. We illustrate the usefulness of caspo by revisiting a model of pro-growth and inflammatory pathways in liver cells. We show that, if experimental error is taken into account, there are thousands (11 700) of models compatible with the data. Despite the large number, we can extract structural features from the models, such as links that are always (or never) present or modules that appear in a mutual exclusive fashion. To further characterize this family of models, we investigate the input-output behavior of the models. We find 91 behaviors across the 11 700 models and we suggest new experiments to discriminate among them. Our results underscore the importance of characterizing in a global and exhaustive manner the family of feasible models, with important implications for experimental design. caspo is freely available for download (license GPLv3) and as a web service at http://caspo.genouest.org/. Supplementary materials are available at Bioinformatics online. santiago.videla@irisa.fr.

On the use of harmony search algorithm in the training of wavelet neural networks

Science.gov (United States)

Lai, Kee Huong; Zainuddin, Zarita; Ong, Pauline

2015-10-01

Wavelet neural networks (WNNs) are a class of feedforward neural networks that have been used in a wide range of industrial and engineering applications to model the complex relationships between the given inputs and outputs. The training of WNNs involves the configuration of the weight values between neurons. The backpropagation training algorithm, which is a gradient-descent method, can be used for this training purpose. Nonetheless, the solutions found by this algorithm often get trapped at local minima. In this paper, a harmony search-based algorithm is proposed for the training of WNNs. The training of WNNs, thus can be formulated as a continuous optimization problem, where the objective is to maximize the overall classification accuracy. Each candidate solution proposed by the harmony search algorithm represents a specific WNN architecture. In order to speed up the training process, the solution space is divided into disjoint partitions during the random initialization step of harmony search algorithm. The proposed training algorithm is tested onthree benchmark problems from the UCI machine learning repository, as well as one real life application, namely, the classification of electroencephalography signals in the task of epileptic seizure detection. The results obtained show that the proposed algorithm outperforms the traditional harmony search algorithm in terms of overall classification accuracy.

Foreign Language Optical Character Recognition, Phase II: Arabic and Persian Training and Test Data Sets

National Research Council Canada - National Science Library

Davidson, Robert

1997-01-01

.... Each data set is divided into a training set, which is made available to developers, and a carefully matched equal-sized set of closely analogous samples, which is reserved for testing of the developers' products...

Manifold absolute pressure estimation using neural network with hybrid training algorithm.

Directory of Open Access Journals (Sweden)

Mohd Taufiq Muslim

Full Text Available In a modern small gasoline engine fuel injection system, the load of the engine is estimated based on the measurement of the manifold absolute pressure (MAP sensor, which took place in the intake manifold. This paper present a more economical approach on estimating the MAP by using only the measurements of the throttle position and engine speed, resulting in lower implementation cost. The estimation was done via two-stage multilayer feed-forward neural network by combining Levenberg-Marquardt (LM algorithm, Bayesian Regularization (BR algorithm and Particle Swarm Optimization (PSO algorithm. Based on the results found in 20 runs, the second variant of the hybrid algorithm yields a better network performance than the first variant of hybrid algorithm, LM, LM with BR and PSO by estimating the MAP closely to the simulated MAP values. By using a valid experimental training data, the estimator network that trained with the second variant of the hybrid algorithm showed the best performance among other algorithms when used in an actual retrofit fuel injection system (RFIS. The performance of the estimator was also validated in steady-state and transient condition by showing a closer MAP estimation to the actual value.

Manifold absolute pressure estimation using neural network with hybrid training algorithm.

Science.gov (United States)

Muslim, Mohd Taufiq; Selamat, Hazlina; Alimin, Ahmad Jais; Haniff, Mohamad Fadzli

2017-01-01

In a modern small gasoline engine fuel injection system, the load of the engine is estimated based on the measurement of the manifold absolute pressure (MAP) sensor, which took place in the intake manifold. This paper present a more economical approach on estimating the MAP by using only the measurements of the throttle position and engine speed, resulting in lower implementation cost. The estimation was done via two-stage multilayer feed-forward neural network by combining Levenberg-Marquardt (LM) algorithm, Bayesian Regularization (BR) algorithm and Particle Swarm Optimization (PSO) algorithm. Based on the results found in 20 runs, the second variant of the hybrid algorithm yields a better network performance than the first variant of hybrid algorithm, LM, LM with BR and PSO by estimating the MAP closely to the simulated MAP values. By using a valid experimental training data, the estimator network that trained with the second variant of the hybrid algorithm showed the best performance among other algorithms when used in an actual retrofit fuel injection system (RFIS). The performance of the estimator was also validated in steady-state and transient condition by showing a closer MAP estimation to the actual value.

Working memory training in congenitally blind individuals results in an integration of occipital cortex in functional networks.

Science.gov (United States)

Gudi-Mindermann, Helene; Rimmele, Johanna M; Nolte, Guido; Bruns, Patrick; Engel, Andreas K; Röder, Brigitte

2018-08-01

The functional relevance of crossmodal activation (e.g. auditory activation of occipital brain regions) in congenitally blind individuals is still not fully understood. The present study tested whether the occipital cortex of blind individuals is integrated into a challenged functional network. A working memory (WM) training over four sessions was implemented. Congenitally blind and matched sighted participants were adaptively trained with an n-back task employing either voices (auditory training) or tactile stimuli (tactile training). In addition, a minimally demanding 1-back task served as an active control condition. Power and functional connectivity of EEG activity evolving during the maintenance period of an auditory 2-back task were analyzed, run prior to and after the WM training. Modality-specific (following auditory training) and modality-independent WM training effects (following both auditory and tactile training) were assessed. Improvements in auditory WM were observed in all groups, and blind and sighted individuals did not differ in training gains. Auditory and tactile training of sighted participants led, relative to the active control group, to an increase in fronto-parietal theta-band power, suggesting a training-induced strengthening of the existing modality-independent WM network. No power effects were observed in the blind. Rather, after auditory training the blind showed a decrease in theta-band connectivity between central, parietal, and occipital electrodes compared to the blind tactile training and active control groups. Furthermore, in the blind auditory training increased beta-band connectivity between fronto-parietal, central and occipital electrodes. In the congenitally blind, these findings suggest a stronger integration of occipital areas into the auditory WM network. Copyright © 2018 Elsevier B.V. All rights reserved.

Let's Wiggle with 5-2-1-0: Curriculum Development for Training Childcare Providers to Promote Activity in Childcare Settings.

Science.gov (United States)

Vinci, Debra M; Whitt-Glover, Melicia C; Wirth, Christopher K; Kraus, Caroline; Venezia, Alexandra P

2016-01-01

Overweight and obesity are increasing in preschool children in the US. Policy, systems, and environmental change interventions in childcare settings can improve obesity-related behaviors. The aim of this study was to develop and pilot an intervention to train childcare providers to promote physical activity (PA) in childcare classrooms. An evidence scan, key informant (n = 34) and focus group (n = 20) interviews with childcare directors and staff, and environmental self-assessment of childcare facilities (n = 22) informed the design of the training curriculum. Feedback from the interviews indicated that childcare providers believed in the importance of teaching children about PA and were supportive of training teachers to incorporate PA into classroom settings. The Promoting Physical Activity in Childcare Setting Curriculum was developed and training was implemented with 16 teachers. Participants reported a positive experience with the hands-on training and reported acquiring new knowledge that they intended to implement in their childcare settings. Our findings highlight the feasibility of working with childcare staff to develop PA training and curriculum. Next steps include evaluating the curriculum in additional childcare settings and childcare staff implementation of the curriculum to understand the effectiveness of the training on PA levels of children.

End-to-End Delay Model for Train Messaging over Public Land Mobile Networks

Directory of Open Access Journals (Sweden)

Franco Mazzenga

2017-11-01

Full Text Available Modern train control systems rely on a dedicated radio network for train to ground communications. A number of possible alternatives have been analysed to adopt the European Rail Traffic Management System/European Train Control System (ERTMS/ETCS control system on local/regional lines to improve transport capacity. Among them, a communication system based on public networks (cellular&satellite provides an interesting, effective and alternative solution to proprietary and expensive radio networks. To analyse performance of this solution, it is necessary to model the end-to-end delay and message loss to fully characterize the message transfer process from train to ground and vice versa. Starting from the results of a railway test campaign over a 300 km railway line for a cumulative 12,000 traveled km in 21 days, in this paper, we derive a statistical model for the end-to-end delay required for delivering messages. In particular, we propose a two states model allowing for reproducing the main behavioral characteristics of the end-to-end delay as observed experimentally. Model formulation has been derived after deep analysis of the recorded experimental data. When it is applied to model a realistic scenario, it allows for explicitly accounting for radio coverage characteristics, the received power level, the handover points along the line and for the serving radio technology. As an example, the proposed model is used to generate the end-to-end delay profile in a realistic scenario.

SAGRAD: A Program for Neural Network Training with Simulated Annealing and the Conjugate Gradient Method.

Science.gov (United States)

Bernal, Javier; Torres-Jimenez, Jose

2015-01-01

SAGRAD (Simulated Annealing GRADient), a Fortran 77 program for computing neural networks for classification using batch learning, is discussed. Neural network training in SAGRAD is based on a combination of simulated annealing and Møller's scaled conjugate gradient algorithm, the latter a variation of the traditional conjugate gradient method, better suited for the nonquadratic nature of neural networks. Different aspects of the implementation of the training process in SAGRAD are discussed, such as the efficient computation of gradients and multiplication of vectors by Hessian matrices that are required by Møller's algorithm; the (re)initialization of weights with simulated annealing required to (re)start Møller's algorithm the first time and each time thereafter that it shows insufficient progress in reaching a possibly local minimum; and the use of simulated annealing when Møller's algorithm, after possibly making considerable progress, becomes stuck at a local minimum or flat area of weight space. Outlines of the scaled conjugate gradient algorithm, the simulated annealing procedure and the training process used in SAGRAD are presented together with results from running SAGRAD on two examples of training data.

A Regularizer Approach for RBF Networks Under the Concurrent Weight Failure Situation.

Science.gov (United States)

Leung, Chi-Sing; Wan, Wai Yan; Feng, Ruibin

2017-06-01

Many existing results on fault-tolerant algorithms focus on the single fault source situation, where a trained network is affected by one kind of weight failure. In fact, a trained network may be affected by multiple kinds of weight failure. This paper first studies how the open weight fault and the multiplicative weight noise degrade the performance of radial basis function (RBF) networks. Afterward, we define the objective function for training fault-tolerant RBF networks. Based on the objective function, we then develop two learning algorithms, one batch mode and one online mode. Besides, the convergent conditions of our online algorithm are investigated. Finally, we develop a formula to estimate the test set error of faulty networks trained from our approach. This formula helps us to optimize some tuning parameters, such as RBF width.

Modeling of quasistatic magnetic hysteresis with feed-forward neural networks

International Nuclear Information System (INIS)

Makaveev, Dimitre; Dupre, Luc; De Wulf, Marc; Melkebeek, Jan

2001-01-01

A modeling technique for rate-independent (quasistatic) scalar magnetic hysteresis is presented, using neural networks. Based on the theory of dynamic systems and the wiping-out and congruency properties of the classical scalar Preisach hysteresis model, the choice of a feed-forward neural network model is motivated. The neural network input parameters at each time step are the corresponding magnetic field strength and memory state, thereby assuring accurate prediction of the change of magnetic induction. For rate-independent hysteresis, the current memory state can be determined by the last extreme magnetic field strength and induction values, kept in memory. The choice of a network training set is motivated and the performance of the network is illustrated for a test set not used during training. Very accurate prediction of both major and minor hysteresis loops is observed, proving that the neural network technique is suitable for hysteresis modeling. [copyright] 2001 American Institute of Physics

The Train Driver Recovery Problem - a Set Partitioning Based Model and Solution Method

DEFF Research Database (Denmark)

Rezanova, Natalia Jurjevna; Ryan, David

2010-01-01

The need to recover a train driver schedule occurs during major disruptions in the daily railway operations. Based on data from the Danish passenger railway operator DSB S-tog A/S, a solution method to the train driver recovery problem (TDRP) is developed. The TDRP is formulated as a set...... branching strategy using the depth-first search of the Branch & Bound tree. The LP relaxation of the TDRP possesses strong integer properties. We present test scenarios generated from the historical real-life operations data of DSB S-tog A/S. The numerical results show that all but one tested instances...... partitioning problem. We define a disruption neighbourhood by identifying a small set of drivers and train tasks directly affected by the disruption. Based on the disruption neighbourhood, the TDRP model is formed and solved. If the TDRP solution provides a feasible recovery for the drivers within...

Improvement of training set structure in fusion data cleaning using Time-Domain Global Similarity method

International Nuclear Information System (INIS)

Liu, J.; Lan, T.; Qin, H.

2017-01-01

Traditional data cleaning identifies dirty data by classifying original data sequences, which is a class-imbalanced problem since the proportion of incorrect data is much less than the proportion of correct ones for most diagnostic systems in Magnetic Confinement Fusion (MCF) devices. When using machine learning algorithms to classify diagnostic data based on class-imbalanced training set, most classifiers are biased towards the major class and show very poor classification rates on the minor class. By transforming the direct classification problem about original data sequences into a classification problem about the physical similarity between data sequences, the class-balanced effect of Time-Domain Global Similarity (TDGS) method on training set structure is investigated in this paper. Meanwhile, the impact of improved training set structure on data cleaning performance of TDGS method is demonstrated with an application example in EAST POlarimetry-INTerferometry (POINT) system.

Training Deep Convolutional Neural Networks with Resistive Cross-Point Devices

Directory of Open Access Journals (Sweden)

Tayfun Gokmen

2017-10-01

Full Text Available In a previous work we have detailed the requirements for obtaining maximal deep learning performance benefit by implementing fully connected deep neural networks (DNN in the form of arrays of resistive devices. Here we extend the concept of Resistive Processing Unit (RPU devices to convolutional neural networks (CNNs. We show how to map the convolutional layers to fully connected RPU arrays such that the parallelism of the hardware can be fully utilized in all three cycles of the backpropagation algorithm. We find that the noise and bound limitations imposed by the analog nature of the computations performed on the arrays significantly affect the training accuracy of the CNNs. Noise and bound management techniques are presented that mitigate these problems without introducing any additional complexity in the analog circuits and that can be addressed by the digital circuits. In addition, we discuss digitally programmable update management and device variability reduction techniques that can be used selectively for some of the layers in a CNN. We show that a combination of all those techniques enables a successful application of the RPU concept for training CNNs. The techniques discussed here are more general and can be applied beyond CNN architectures and therefore enables applicability of the RPU approach to a large class of neural network architectures.

Performance of a visuomotor walking task in an augmented reality training setting.

Science.gov (United States)

Haarman, Juliet A M; Choi, Julia T; Buurke, Jaap H; Rietman, Johan S; Reenalda, Jasper

2017-12-01

Visual cues can be used to train walking patterns. Here, we studied the performance and learning capacities of healthy subjects executing a high-precision visuomotor walking task, in an augmented reality training set-up. A beamer was used to project visual stepping targets on the walking surface of an instrumented treadmill. Two speeds were used to manipulate task difficulty. All participants (n = 20) had to change their step length to hit visual stepping targets with a specific part of their foot, while walking on a treadmill over seven consecutive training blocks, each block composed of 100 stepping targets. Distance between stepping targets was varied between short, medium and long steps. Training blocks could either be composed of random stepping targets (no fixed sequence was present in the distance between the stepping targets) or sequenced stepping targets (repeating fixed sequence was present). Random training blocks were used to measure non-specific learning and sequenced training blocks were used to measure sequence-specific learning. Primary outcome measures were performance (% of correct hits), and learning effects (increase in performance over the training blocks: both sequence-specific and non-specific). Secondary outcome measures were the performance and stepping-error in relation to the step length (distance between stepping target). Subjects were able to score 76% and 54% at first try for lower speed (2.3 km/h) and higher speed (3.3 km/h) trials, respectively. Performance scores did not increase over the course of the trials, nor did the subjects show the ability to learn a sequenced walking task. Subjects were better able to hit targets while increasing their step length, compared to shortening it. In conclusion, augmented reality training by use of the current set-up was intuitive for the user. Suboptimal feedback presentation might have limited the learning effects of the subjects. Copyright © 2017 Elsevier B.V. All rights reserved.

Strong Relaxations for the Train Timetabling Problem Using Connected Configurations

OpenAIRE

Fischer, Frank; Schlechte, Thomas

2017-01-01

The task of the train timetabling problem or track allocation problem is to find conflict free schedules for a set of trains with predefined routes in a railway network. Especially for non-periodic instances models based on time expanded networks are often used. Unfortunately, the linear programming relaxation of these models is often extremely weak because these models do not describe combinatorial relations like overtaking possibilities very well. In this paper we extend the model by so cal...

Cooperative VET in Training Networks: Analysing the Free-Rider Problem in a Sociology-of-Conventions Perspective

Directory of Open Access Journals (Sweden)

Regula Julia Leemann

2015-12-01

Full Text Available In training networks, particularly small and medium-sized enterprises pool their resources to train apprentices within the framework of the dual VET system, while an intermediary organisation is tasked with managing operations. Over the course of their apprenticeship, the apprentices switch from one training company to another on a (half- yearly basis. Drawing on a case study of four training networks in Switzerland and the theoretical framework of the sociology of conventions, this paper aims to understand the reasons for the slow dissemination and reluctant adoption of this promising form of organising VET in Switzerland. The results of the study show that the system of moving from one company to another creages a variety of free-rider constellations in the distribution of the collectively generated corporative benefits. This explains why companies are reluctant to participate in this model. For the network to be sustainable, the intermediary organisation has to address discontent arising from free-rider problems while taking into account that the solutions found are always tentative and will often result in new free-rider problems.

Self-Consistent Scheme for Spike-Train Power Spectra in Heterogeneous Sparse Networks.

Science.gov (United States)

Pena, Rodrigo F O; Vellmer, Sebastian; Bernardi, Davide; Roque, Antonio C; Lindner, Benjamin

2018-01-01

Recurrent networks of spiking neurons can be in an asynchronous state characterized by low or absent cross-correlations and spike statistics which resemble those of cortical neurons. Although spatial correlations are negligible in this state, neurons can show pronounced temporal correlations in their spike trains that can be quantified by the autocorrelation function or the spike-train power spectrum. Depending on cellular and network parameters, correlations display diverse patterns (ranging from simple refractory-period effects and stochastic oscillations to slow fluctuations) and it is generally not well-understood how these dependencies come about. Previous work has explored how the single-cell correlations in a homogeneous network (excitatory and inhibitory integrate-and-fire neurons with nearly balanced mean recurrent input) can be determined numerically from an iterative single-neuron simulation. Such a scheme is based on the fact that every neuron is driven by the network noise (i.e., the input currents from all its presynaptic partners) but also contributes to the network noise, leading to a self-consistency condition for the input and output spectra. Here we first extend this scheme to homogeneous networks with strong recurrent inhibition and a synaptic filter, in which instabilities of the previous scheme are avoided by an averaging procedure. We then extend the scheme to heterogeneous networks in which (i) different neural subpopulations (e.g., excitatory and inhibitory neurons) have different cellular or connectivity parameters; (ii) the number and strength of the input connections are random (Erdős-Rényi topology) and thus different among neurons. In all heterogeneous cases, neurons are lumped in different classes each of which is represented by a single neuron in the iterative scheme; in addition, we make a Gaussian approximation of the input current to the neuron. These approximations seem to be justified over a broad range of parameters as

Self-Consistent Scheme for Spike-Train Power Spectra in Heterogeneous Sparse Networks

Directory of Open Access Journals (Sweden)

Rodrigo F. O. Pena

2018-03-01

Full Text Available Recurrent networks of spiking neurons can be in an asynchronous state characterized by low or absent cross-correlations and spike statistics which resemble those of cortical neurons. Although spatial correlations are negligible in this state, neurons can show pronounced temporal correlations in their spike trains that can be quantified by the autocorrelation function or the spike-train power spectrum. Depending on cellular and network parameters, correlations display diverse patterns (ranging from simple refractory-period effects and stochastic oscillations to slow fluctuations and it is generally not well-understood how these dependencies come about. Previous work has explored how the single-cell correlations in a homogeneous network (excitatory and inhibitory integrate-and-fire neurons with nearly balanced mean recurrent input can be determined numerically from an iterative single-neuron simulation. Such a scheme is based on the fact that every neuron is driven by the network noise (i.e., the input currents from all its presynaptic partners but also contributes to the network noise, leading to a self-consistency condition for the input and output spectra. Here we first extend this scheme to homogeneous networks with strong recurrent inhibition and a synaptic filter, in which instabilities of the previous scheme are avoided by an averaging procedure. We then extend the scheme to heterogeneous networks in which (i different neural subpopulations (e.g., excitatory and inhibitory neurons have different cellular or connectivity parameters; (ii the number and strength of the input connections are random (Erdős-Rényi topology and thus different among neurons. In all heterogeneous cases, neurons are lumped in different classes each of which is represented by a single neuron in the iterative scheme; in addition, we make a Gaussian approximation of the input current to the neuron. These approximations seem to be justified over a broad range of

Clustering and training set selection methods for improving the accuracy of quantitative laser induced breakdown spectroscopy

International Nuclear Information System (INIS)

Anderson, Ryan B.; Bell, James F.; Wiens, Roger C.; Morris, Richard V.; Clegg, Samuel M.

2012-01-01

We investigated five clustering and training set selection methods to improve the accuracy of quantitative chemical analysis of geologic samples by laser induced breakdown spectroscopy (LIBS) using partial least squares (PLS) regression. The LIBS spectra were previously acquired for 195 rock slabs and 31 pressed powder geostandards under 7 Torr CO 2 at a stand-off distance of 7 m at 17 mJ per pulse to simulate the operational conditions of the ChemCam LIBS instrument on the Mars Science Laboratory Curiosity rover. The clustering and training set selection methods, which do not require prior knowledge of the chemical composition of the test-set samples, are based on grouping similar spectra and selecting appropriate training spectra for the partial least squares (PLS2) model. These methods were: (1) hierarchical clustering of the full set of training spectra and selection of a subset for use in training; (2) k-means clustering of all spectra and generation of PLS2 models based on the training samples within each cluster; (3) iterative use of PLS2 to predict sample composition and k-means clustering of the predicted compositions to subdivide the groups of spectra; (4) soft independent modeling of class analogy (SIMCA) classification of spectra, and generation of PLS2 models based on the training samples within each class; (5) use of Bayesian information criteria (BIC) to determine an optimal number of clusters and generation of PLS2 models based on the training samples within each cluster. The iterative method and the k-means method using 5 clusters showed the best performance, improving the absolute quadrature root mean squared error (RMSE) by ∼ 3 wt.%. The statistical significance of these improvements was ∼ 85%. Our results show that although clustering methods can modestly improve results, a large and diverse training set is the most reliable way to improve the accuracy of quantitative LIBS. In particular, additional sulfate standards and specifically

Higher surgical training opportunities in the general hospital setting; getting the balance right.

Science.gov (United States)

Robertson, I; Traynor, O; Khan, W; Waldron, R; Barry, K

2013-12-01

The general hospital can play an important role in training of higher surgical trainees (HSTs) in Ireland and abroad. Training opportunities in such a setting have not been closely analysed to date. The aim of this study was to quantify operative exposure for HSTs over a 5-year period in a single institution. Analysis of electronic training logbooks (over a 5-year period, 2007-2012) was performed for general surgery trainees on the higher surgical training programme in Ireland. The most commonly performed adult and paediatric procedures per trainee, per year were analysed. Standard general surgery operations such as herniae (average 58, range 32-86) and cholecystectomy (average 60, range 49-72) ranked highly in each logbook. The most frequently performed emergency operations were appendicectomy (average 45, range 33-53) and laparotomy for acute abdomen (average 48, range 10-79). Paediatric surgical experience included appendicectomy, circumcision, orchidopexy and hernia/hydrocoele repair. Overall, the procedure most commonly performed in the adult setting was endoscopy, with each trainee recording an average of 116 (range 98-132) oesophagogastroduodenoscopies and 284 (range 227-354) colonoscopies. General hospitals continue to play a major role in the training of higher surgical trainees. Analysis of the electronic logbooks over a 5-year period reveals the high volume of procedures available to trainees in a non-specialist centre. Such training opportunities are invaluable in the context of changing work practices and limited resources.

Monitoring training response in young Friesian dressage horses using two different standardised exercise tests (SETs)

NARCIS (Netherlands)

de Bruijn, Cornelis Marinus; Houterman, Willem; Ploeg, Margreet; Ducro, Bart; Boshuizen, Berit; Goethals, Klaartje; Verdegaal, Elisabeth-Lidwien; Delesalle, Catherine

2017-01-01

BACKGROUND: Most Friesian horses reach their anaerobic threshold during a standardized exercise test (SET) which requires lower intensity exercise than daily routine training. AIM: to study strengths and weaknesses of an alternative SET-protocol. Two different SETs (SETA and SETB) were applied

Monitoring training response in young Friesian dressage horses using two different standardised exercise tests (SETs)

NARCIS (Netherlands)

Bruijn, de Cornelis Marinus; Houterman, Willem; Ploeg, Margreet; Ducro, Bart; Boshuizen, Berit; Goethals, Klaartje; Verdegaal, Elisabeth Lidwien; Delesalle, Catherine

2017-01-01

Background: Most Friesian horses reach their anaerobic threshold during a standardized exercise test (SET) which requires lower intensity exercise than daily routine training. Aim: to study strengths and weaknesses of an alternative SET-protocol. Two different SETs (SETA and SETB) were applied

Diagnostics of Nuclear Reactor Accidents Based on Particle Swarm Optimization Trained Neural Networks

International Nuclear Information System (INIS)

Abdel-Aal, M.M.Z.

2004-01-01

Automation in large, complex systems such as chemical plants, electrical power generation, aerospace and nuclear plants has been steadily increasing in the recent past. automated diagnosis and control forms a necessary part of these systems,this contains thousands of alarms processing in every component, subsystem and system. so the accurate and speed of diagnosis of faults is an important factors in operation and maintaining their health and continued operation and in reducing of repair and recovery time. using of artificial intelligence facilitates the alarm classifications and faults diagnosis to control any abnormal events during the operation cycle of the plant. thesis work uses the artificial neural network as a powerful classification tool. the work basically is has two components, the first is to effectively train the neural network using particle swarm optimization, which non-derivative based technique. to achieve proper training of the neural network to fault classification problem and comparing this technique to already existing techniques

Computationally Efficient Neural Network Intrusion Security Awareness

Energy Technology Data Exchange (ETDEWEB)

Todd Vollmer; Milos Manic

2009-08-01

An enhanced version of an algorithm to provide anomaly based intrusion detection alerts for cyber security state awareness is detailed. A unique aspect is the training of an error back-propagation neural network with intrusion detection rule features to provide a recognition basis. Network packet details are subsequently provided to the trained network to produce a classification. This leverages rule knowledge sets to produce classifications for anomaly based systems. Several test cases executed on ICMP protocol revealed a 60% identification rate of true positives. This rate matched the previous work, but 70% less memory was used and the run time was reduced to less than 1 second from 37 seconds.

Development of Artificial Neural Network Model of Crude Oil Distillation Column

Directory of Open Access Journals (Sweden)

Ali Hussein Khalaf

2016-02-01

Full Text Available Artificial neural network in MATLAB simulator is used to model Baiji crude oil distillation unit based on data generated from aspen-HYSYS simulator. Thirteen inputs, six outputs and over 1487 data set are used to model the actual unit. Nonlinear autoregressive network with exogenous inputs (NARXand back propagation algorithm are used for training. Seventy percent of data are used for training the network while the remaining  thirty percent are used for testing  and validating the network to determine its prediction accuracy. One hidden layer and 34 hidden neurons are used for the proposed network with MSE of 0.25 is obtained. The number of neuron are selected based on less MSE for the network. The model founded to predict the optimal operating conditions for different objective functions within the training limit since ANN models are poor extrapolators. They are usually only reliable within the range of data that they had been trained for.

Development of Artificial Neural Network Model of Crude Oil Distillation Column

Directory of Open Access Journals (Sweden)

Duraid F. Ahmed

2016-02-01

Full Text Available Artificial neural network in MATLAB simulator is used to model Baiji crude oil distillation unit based on data generated from aspen-HYSYS simulator. Thirteen inputs, six outputs and over 1487 data set are used to model the actual unit. Nonlinear autoregressive network with exogenous inputs (NARX and back propagation algorithm are used for training. Seventy percent of data are used for training the network while the remaining thirty percent are used for testing and validating the network to determine its prediction accuracy. One hidden layer and 34 hidden neurons are used for the proposed network with MSE of 0.25 is obtained. The number of neuron are selected based on less MSE for the network. The model founded to predict the optimal operating conditions for different objective functions within the training limit since ANN models are poor extrapolators. They are usually only reliable within the range of data that they had been trained for.

Advances in Artificial Neural Networks – Methodological Development and Application

Directory of Open Access Journals (Sweden)

Yanbo Huang

2009-08-01

Full Text Available Artificial neural networks as a major soft-computing technology have been extensively studied and applied during the last three decades. Research on backpropagation training algorithms for multilayer perceptron networks has spurred development of other neural network training algorithms for other networks such as radial basis function, recurrent network, feedback network, and unsupervised Kohonen self-organizing network. These networks, especially the multilayer perceptron network with a backpropagation training algorithm, have gained recognition in research and applications in various scientific and engineering areas. In order to accelerate the training process and overcome data over-fitting, research has been conducted to improve the backpropagation algorithm. Further, artificial neural networks have been integrated with other advanced methods such as fuzzy logic and wavelet analysis, to enhance the ability of data interpretation and modeling and to avoid subjectivity in the operation of the training algorithm. In recent years, support vector machines have emerged as a set of high-performance supervised generalized linear classifiers in parallel with artificial neural networks. A review on development history of artificial neural networks is presented and the standard architectures and algorithms of artificial neural networks are described. Furthermore, advanced artificial neural networks will be introduced with support vector machines, and limitations of ANNs will be identified. The future of artificial neural network development in tandem with support vector machines will be discussed in conjunction with further applications to food science and engineering, soil and water relationship for crop management, and decision support for precision agriculture. Along with the network structures and training algorithms, the applications of artificial neural networks will be reviewed as well, especially in the fields of agricultural and biological

Photo-z with CuBANz: An improved photometric redshift estimator using Clustering aided Back propagation Neural network

Science.gov (United States)

Samui, Saumyadip; Samui Pal, Shanoli

2017-02-01

We present an improved photometric redshift estimator code, CuBANz, that is publicly available at https://goo.gl/fpk90V. It uses the back propagation neural network along with clustering of the training set, which makes it more efficient than existing neural network codes. In CuBANz, the training set is divided into several self learning clusters with galaxies having similar photometric properties and spectroscopic redshifts within a given span. The clustering algorithm uses the color information (i.e. u - g , g - r etc.) rather than the apparent magnitudes at various photometric bands as the photometric redshift is more sensitive to the flux differences between different bands rather than the actual values. Separate neural networks are trained for each cluster using all possible colors, magnitudes and uncertainties in the measurements. For a galaxy with unknown redshift, we identify the closest possible clusters having similar photometric properties and use those clusters to get the photometric redshifts using the particular networks that were trained using those cluster members. For galaxies that do not match with any training cluster, the photometric redshifts are obtained from a separate network that uses entire training set. This clustering method enables us to determine the redshifts more accurately. SDSS Stripe 82 catalog has been used here for the demonstration of the code. For the clustered sources with redshift range zspec training/testing phase is as low as 0.03 compared to the existing ANNz code that provides residual error on the same test data set of 0.05. Further, we provide a much better estimate of the uncertainty of the derived photometric redshift.

Neural network modelling of planform geometry of headland-bay beaches

Science.gov (United States)

Iglesias, G.; López, I.; Castro, A.; Carballo, R.

2009-02-01

The shoreline of beaches in the lee of coastal salients or man-made structures, usually known as headland-bay beaches, has a distinctive curvature; wave fronts curve as a result of wave diffraction at the headland and in turn cause the shoreline to bend. The ensuing curved planform is of great interest both as a peculiar landform and in the context of engineering projects in which it is necessary to predict how a coastal structure will affect the sandy shoreline in its lee. A number of empirical models have been put forward, each based on a specific equation. A novel approach, based on the application of artificial neural networks, is presented in this work. Unlike the conventional method, no particular equation of the planform is embedded in the model. Instead, it is the model itself that learns about the problem from a series of examples of headland-bay beaches (the training set) and thereafter applies this self-acquired knowledge to other cases (the test set) for validation. Twenty-three headland-bay beaches from around the world were selected, of which sixteen and seven make up the training and test sets, respectively. As there is no well-developed theory for deciding upon the most convenient neural network architecture to deal with a particular data set, an experimental study was conducted in which ten different architectures with one and two hidden neuron layers and five training algorithms - 50 different options combining network architecture and training algorithm - were compared. Each of these options was implemented, trained and tested in order to find the best-performing approach for modelling the planform of headland-bay beaches. Finally, the selected neural network model was compared with a state-of-the-art planform model and was shown to outperform it.

Cooperative VET in Training Networks: Analysing the Free-Rider Problem in a Sociology-of-Conventions Perspective

Science.gov (United States)

Leemann, Regula Julia; Imdorf, Christian

2015-01-01

In training networks, particularly small and medium-sized enterprises pool their resources to train apprentices within the framework of the dual VET system, while an intermediary organisation is tasked with managing operations. Over the course of their apprenticeship, the apprentices switch from one training company to another on a (half-) yearly…

Role of ventilation scintigraphy in diagnosis of acute pulmonary embolism: an evaluation using artificial neural networks

International Nuclear Information System (INIS)

Evander, Eva; Holst, Holger; Jaerund, Andreas; Wollmer, Per; Edenbrandt, Lars; Ohlsson, Mattias; Aastroem, Karl

2003-01-01

The purpose of this study was to assess the value of the ventilation study in the diagnosis of acute pulmonary embolism using a new automated method. Either perfusion scintigrams alone or two different combinations of ventilation/perfusion scintigrams were used as the only source of information regarding pulmonary embolism. A completely automated method based on computerised image processing and artificial neural networks was used for the interpretation. Three artificial neural networks were trained for the diagnosis of pulmonary embolism. Each network was trained with 18 automatically obtained features. Three different sets of features originating from three sets of scintigrams were used. One network was trained using features obtained from each set of perfusion scintigrams, including six projections. The second network was trained using features from each set of (joint) ventilation and perfusion studies in six projections. A third network was trained using features from the perfusion study in six projections combined with a single ventilation image from the posterior view. A total of 1,087 scintigrams from patients with suspected pulmonary embolism were used for network training. The test group consisted of 102 patients who had undergone both scintigraphy and pulmonary angiography. Performances in the test group were measured as area under the receiver operation characteristic curve. The performance of the neural network in interpreting perfusion scintigrams alone was 0.79 (95% confidence limits 0.71-0.86). When one ventilation image (posterior view) was added to the perfusion study, the performance was 0.84 (0.77-0.90). This increase was statistically significant (P=0.022). The performance increased to 0.87 (0.81-0.93) when all perfusion and ventilation images were used, and the increase in performance from 0.79 to 0.87 was also statistically significant (P=0.016). The automated method presented here for the interpretation of lung scintigrams shows a significant

Generative adversarial networks for brain lesion detection

Science.gov (United States)

Alex, Varghese; Safwan, K. P. Mohammed; Chennamsetty, Sai Saketh; Krishnamurthi, Ganapathy

2017-02-01

Manual segmentation of brain lesions from Magnetic Resonance Images (MRI) is cumbersome and introduces errors due to inter-rater variability. This paper introduces a semi-supervised technique for detection of brain lesion from MRI using Generative Adversarial Networks (GANs). GANs comprises of a Generator network and a Discriminator network which are trained simultaneously with the objective of one bettering the other. The networks were trained using non lesion patches (n=13,000) from 4 different MR sequences. The network was trained on BraTS dataset and patches were extracted from regions excluding tumor region. The Generator network generates data by modeling the underlying probability distribution of the training data, (PData). The Discriminator learns the posterior probability P (Label Data) by classifying training data and generated data as "Real" or "Fake" respectively. The Generator upon learning the joint distribution, produces images/patches such that the performance of the Discriminator on them are random, i.e. P (Label Data = GeneratedData) = 0.5. During testing, the Discriminator assigns posterior probability values close to 0.5 for patches from non lesion regions, while patches centered on lesion arise from a different distribution (PLesion) and hence are assigned lower posterior probability value by the Discriminator. On the test set (n=14), the proposed technique achieves whole tumor dice score of 0.69, sensitivity of 91% and specificity of 59%. Additionally the generator network was capable of generating non lesion patches from various MR sequences.

Classification of urine sediment based on convolution neural network

Science.gov (United States)

Pan, Jingjing; Jiang, Cunbo; Zhu, Tiantian

2018-04-01

By designing a new convolution neural network framework, this paper breaks the constraints of the original convolution neural network framework requiring large training samples and samples of the same size. Move and cropping the input images, generate the same size of the sub-graph. And then, the generated sub-graph uses the method of dropout, increasing the diversity of samples and preventing the fitting generation. Randomly select some proper subset in the sub-graphic set and ensure that the number of elements in the proper subset is same and the proper subset is not the same. The proper subsets are used as input layers for the convolution neural network. Through the convolution layer, the pooling, the full connection layer and output layer, we can obtained the classification loss rate of test set and training set. In the red blood cells, white blood cells, calcium oxalate crystallization classification experiment, the classification accuracy rate of 97% or more.

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

Functional MRI neurofeedback training on connectivity between two regions induces long-lasting changes in intrinsic functional network

Directory of Open Access Journals (Sweden)

Fukuda eMegumi

2015-03-01

Full Text Available Motor or perceptual learning is known to influence functional connectivity between brain regions and induce short-term changes in the intrinsic functional networks revealed as correlations in slow blood-oxygen-level dependent (BOLD signal fluctuations. However, no cause-and-effect relationship has been elucidated between a specific change in connectivity and a long-term change in global networks. Here, we examine the hypothesis that functional connectivity (i.e. temporal correlation between two regions is increased and preserved for a long time when two regions are simultaneously activated or deactivated. Using the connectivity-neurofeedback training paradigm, subjects successfully learned to increase the correlation of activity between the lateral parietal and primary motor areas, regions that belong to different intrinsic networks and negatively correlated before training under the resting conditions. Furthermore, whole-brain hypothesis-free analysis as well as functional network analyses demonstrated that the correlation in the resting state between these areas as well as the correlation between the intrinsic networks that include the areas increased for at least two months. These findings indicate that the connectivity-neurofeedback training can cause long-term changes in intrinsic connectivity and that intrinsic networks can be shaped by experience-driven modulation of regional correlation.

Individual Channel Estimation in a Diamond Relay Network Using Relay-Assisted Training

Directory of Open Access Journals (Sweden)

Xianwen He

2017-01-01

Full Text Available We consider the training design and channel estimation in the amplify-and-forward (AF diamond relay network. Our strategy is to transmit the source training in time-multiplexing (TM mode while each relay node superimposes its own relay training over the amplified received data signal without bandwidth expansion. The principal challenge is to obtain accurate channel state information (CSI of second-hop link due to the multiaccess interference (MAI and cooperative data interference (CDI. To maintain the orthogonality between data and training, a modified relay-assisted training scheme is proposed to migrate the CDI, where some of the cooperative data at the relay are discarded to accommodate relay training. Meanwhile, a couple of optimal zero-correlation zone (ZCZ relay-assisted sequences are designed to avoid MAI. At the destination node, the received signals from the two relay nodes are combined to achieve spatial diversity and enhanced data reliability. The simulation results are presented to validate the performance of the proposed schemes.

Clustering and training set selection methods for improving the accuracy of quantitative laser induced breakdown spectroscopy

Energy Technology Data Exchange (ETDEWEB)

Anderson, Ryan B., E-mail: randerson@astro.cornell.edu [Cornell University Department of Astronomy, 406 Space Sciences Building, Ithaca, NY 14853 (United States); Bell, James F., E-mail: Jim.Bell@asu.edu [Arizona State University School of Earth and Space Exploration, Bldg.: INTDS-A, Room: 115B, Box 871404, Tempe, AZ 85287 (United States); Wiens, Roger C., E-mail: rwiens@lanl.gov [Los Alamos National Laboratory, P.O. Box 1663 MS J565, Los Alamos, NM 87545 (United States); Morris, Richard V., E-mail: richard.v.morris@nasa.gov [NASA Johnson Space Center, 2101 NASA Parkway, Houston, TX 77058 (United States); Clegg, Samuel M., E-mail: sclegg@lanl.gov [Los Alamos National Laboratory, P.O. Box 1663 MS J565, Los Alamos, NM 87545 (United States)

2012-04-15

We investigated five clustering and training set selection methods to improve the accuracy of quantitative chemical analysis of geologic samples by laser induced breakdown spectroscopy (LIBS) using partial least squares (PLS) regression. The LIBS spectra were previously acquired for 195 rock slabs and 31 pressed powder geostandards under 7 Torr CO{sub 2} at a stand-off distance of 7 m at 17 mJ per pulse to simulate the operational conditions of the ChemCam LIBS instrument on the Mars Science Laboratory Curiosity rover. The clustering and training set selection methods, which do not require prior knowledge of the chemical composition of the test-set samples, are based on grouping similar spectra and selecting appropriate training spectra for the partial least squares (PLS2) model. These methods were: (1) hierarchical clustering of the full set of training spectra and selection of a subset for use in training; (2) k-means clustering of all spectra and generation of PLS2 models based on the training samples within each cluster; (3) iterative use of PLS2 to predict sample composition and k-means clustering of the predicted compositions to subdivide the groups of spectra; (4) soft independent modeling of class analogy (SIMCA) classification of spectra, and generation of PLS2 models based on the training samples within each class; (5) use of Bayesian information criteria (BIC) to determine an optimal number of clusters and generation of PLS2 models based on the training samples within each cluster. The iterative method and the k-means method using 5 clusters showed the best performance, improving the absolute quadrature root mean squared error (RMSE) by {approx} 3 wt.%. The statistical significance of these improvements was {approx} 85%. Our results show that although clustering methods can modestly improve results, a large and diverse training set is the most reliable way to improve the accuracy of quantitative LIBS. In particular, additional sulfate standards and

Application of radial basis function neural network to predict soil sorption partition coefficient using topological descriptors.

Science.gov (United States)

Sabour, Mohammad Reza; Moftakhari Anasori Movahed, Saman

2017-02-01

The soil sorption partition coefficient logK oc is an indispensable parameter that can be used in assessing the environmental risk of organic chemicals. In order to predict soil sorption partition coefficient for different and even unknown compounds in a fast and accurate manner, a radial basis function neural network (RBFNN) model was developed. Eight topological descriptors of 800 organic compounds were used as inputs of the model. These 800 organic compounds were chosen from a large and very diverse data set. Generalized Regression Neural Network (GRNN) was utilized as the function in this neural network model due to its capability to adapt very quickly. Hence, it can be used to predict logK oc for new chemicals, as well. Out of total data set, 560 organic compounds were used for training and 240 to test efficiency of the model. The obtained results indicate that the model performance is very well. The correlation coefficients (R2) for training and test sets were 0.995 and 0.933, respectively. The root-mean square errors (RMSE) were 0.2321 for training set and 0.413 for test set. As the results for both training and test set are extremely satisfactory, the proposed neural network model can be employed not only to predict logK oc of known compounds, but also to be adaptive for prediction of this value precisely for new products that enter the market each year. Copyright © 2016 Elsevier Ltd. All rights reserved.

A managed clinical network for cardiac services: set-up, operation and impact on patient care.

Science.gov (United States)

Stc Hamilton, Karen E; Sullivan, Frank M; Donnan, Peter T; Taylor, Rex; Ikenwilo, Divine; Scott, Anthony; Baker, Chris; Wyke, Sally

2005-01-01

To investigate the set up and operation of a Managed Clinical Network for cardiac services and assess its impact on patient care. This single case study used process evaluation with observational before and after comparison of indicators of quality of care and costs. The study was conducted in Dumfries and Galloway, Scotland and used a three-level framework. Process evaluation of the network set-up and operation through a documentary review of minutes; guidelines and protocols; transcripts of fourteen semi-structured interviews with health service personnel including senior managers, general practitioners, nurses, cardiologists and members of the public. Outcome evaluation of the impact of the network through interrupted time series analysis of clinical data of 202 patients aged less than 76 years admitted to hospital with a confirmed myocardial infarction one-year pre and one-year post, the establishment of the network. The main outcome measures were differences between indicators of quality of care targeted by network protocols. Economic evaluation of the transaction costs of the set-up and operation of the network and the resource costs of the clinical care of the 202 myocardial infarction patients from the time of hospital admission to 6 months post discharge through interrupted time series analysis. The outcome measure was different in National Health Service resource use. Despite early difficulties, the network was successful in bringing together clinicians, patients and managers to redesign services, exhibiting most features of good network management. The role of the energetic lead clinician was crucial, but the network took time to develop and 'bed down'. Its primary "modus operand" was the development of a myocardial infarction pathway and associated protocols. Of sixteen clinical care indicators, two improved significantly following the launch of the network and nine showed improvements, which were not statistically significant. There was no difference

Evaluation and scoring of radiotherapy treatment plans using an artificial neural network

International Nuclear Information System (INIS)

Willoughby, Twyla R.; Starkschall, George; Janjan, Nora A.; Rosen, Isaac I.

1996-01-01

Purpose: The objective of this work was to demonstrate the feasibility of using an artificial neural network to predict the clinical evaluation of radiotherapy treatment plans. Methods and Materials: Approximately 150 treatment plans were developed for 16 patients who received external-beam radiotherapy for soft-tissue sarcomas of the lower extremity. Plans were assigned a figure of merit by a radiation oncologist using a five-point rating scale. Plan scoring was performed by a single physician to ensure consistency in rating. Dose-volume information extracted from a training set of 511 treatment plans on 14 patients was correlated to the physician-generated figure of merit using an artificial neural network. The neural network was tested with a test set of 19 treatment plans on two patients whose plans were not used in the training of the neural net. Results: Physician scoring of treatment plans was consistent to within one point on the rating scale 88% of the time. The neural net reproduced the physician scores in the training set to within one point approximately 90% of the time. It reproduced the physician scores in the test set to within one point approximately 83% of the time. Conclusions: An artificial neural network can be trained to generate a score for a treatment plan that can be correlated to a clinically-based figure of merit. The accuracy of the neural net in scoring plans compares well with the reproducibility of the clinical scoring. The system of radiotherapy treatment plan evaluation using an artificial neural network demonstrates promise as a method for generating a clinically relevant figure of merit

Deconstructing myths, building alliances: a networking model to enhance tobacco control in hospital mental health settings.

Science.gov (United States)

Ballbè, Montse; Gual, Antoni; Nieva, Gemma; Saltó, Esteve; Fernández, Esteve

2016-01-01

Life expectancy for people with severe mental disorders is up to 25 years less in comparison to the general population, mainly due to diseases caused or worsened by smoking. However, smoking is usually a neglected issue in mental healthcare settings. The aim of this article is to describe a strategy to improve tobacco control in the hospital mental healthcare services of Catalonia (Spain). To bridge this gap, the Catalan Network of Smoke-free Hospitals launched a nationwide bottom-up strategy in Catalonia in 2007. The strategy relied on the creation of a working group of key professionals from various hospitals -the early adopters- based on Rogers' theory of the Diffusion of Innovations. In 2016, the working group is composed of professionals from 17 hospitals (70.8% of all hospitals in the region with mental health inpatient units). Since 2007, tobacco control has improved in different areas such as increasing mental health professionals' awareness of smoking, training professionals on smoking cessation interventions and achieving good compliance with the national smoking ban. The working group has produced and disseminated various materials, including clinical practice and best practice guidelines, implemented smoking cessation programmes and organised seminars and training sessions on smoking cessation measures in patients with mental illnesses. The next challenge is to ensure effective follow-up for smoking cessation after discharge. While some areas of tobacco control within these services still require significant improvement, the aforementioned initiative promotes successful tobacco control in these settings. Copyright © 2016 SESPAS. Publicado por Elsevier España, S.L.U. All rights reserved.

Modeling of an ionic polymer metal composite actuator based on an extended Kalman filter trained neural network

International Nuclear Information System (INIS)

Truong, Dinh Quang; Ahn, Kyoung Kwan

2014-01-01

An ion polymer metal composite (IPMC) is an electroactive polymer that bends in response to a small applied electric field as a result of mobility of cations in the polymer network and vice versa. This paper presents an innovative and accurate nonlinear black-box model (NBBM) for estimating the bending behavior of IPMC actuators. The model is constructed via a general multilayer perceptron neural network (GMLPNN) integrated with a smart learning mechanism (SLM) that is based on an extended Kalman filter with self-decoupling ability (SDEKF). Here the GMLPNN is built with an ability to autoadjust its structure based on its characteristic vector. Furthermore, by using the SLM based on the SDEKF, the GMLPNN parameters are optimized with small computational effort, and the modeling accuracy is improved. An apparatus employing an IPMC actuator is first set up to investigate the IPMC characteristics and to generate the data for training and validating the model. The advanced NBBM model for the IPMC system is then created with the proper inputs to estimate IPMC tip displacement. Next, the model is optimized using the SLM mechanism with the training data. Finally, the optimized NBBM model is verified with the validating data. A comparison between this model and the previously developed model is also carried out to prove the effectiveness of the proposed modeling technique. (paper)

An Analysis of Training Focused on Improving SMART Goal Setting for Specific Employee Groups

Science.gov (United States)

Worden, Jeannie M.

2014-01-01

This quantitative study examined the proficiency of employee SMART goal setting following the intervention of employee SMART goal setting training. Current challenges in higher education substantiate the need for employees to align their performance with the mission, vision, and strategic directions of the organization. A performance management…

Voltage regulation in MV networks with dispersed generations by a neural-based multiobjective methodology

Energy Technology Data Exchange (ETDEWEB)

Galdi, Vincenzo [Dipartimento di Ingegneria dell' Informazione e Ingegneria Elettrica, Universita degli studi di Salerno, Via Ponte Don Melillo 1, 84084 Fisciano (Italy); Vaccaro, Alfredo; Villacci, Domenico [Dipartimento di Ingegneria, Universita degli Studi del Sannio, Piazza Roma 21, 82100 Benevento (Italy)

2008-05-15

This paper puts forward the role of learning techniques in addressing the problem of an efficient and optimal centralized voltage control in distribution networks equipped with dispersed generation systems (DGSs). The proposed methodology employs a radial basis function network (RBFN) to identify the multidimensional nonlinear mapping between a vector of observable variables describing the network operating point and the optimal set points of the voltage regulating devices. The RBFN is trained by numerical data generated by solving the voltage regulation problem for a set of network operating points by a rigorous multiobjective solution methodology. The RBFN performance is continuously monitored by a supervisor process that notifies when there is the need of a more accurate solution of the voltage regulation problem if nonoptimal network operating conditions (ex post monitoring) or excessive distances between the actual network state and the neuron's centres (ex ante monitoring) are detected. A more rigorous problem solution, if required, can be obtained by solving the voltage regulation problem by a conventional multiobjective optimization technique. This new solution, in conjunction with the corresponding input vector, is then adopted as a new train data sample to adapt the RBFN. This online training process allows RBFN to (i) adaptively learn the more representative domain space regions of the input/output mapping without needing a prior knowledge of a complete and representative training set, and (ii) manage effectively any time varying phenomena affecting this mapping. The results obtained by simulating the regulation policy in the case of a medium-voltage network are very promising. (author)

Applying long short-term memory recurrent neural networks to intrusion detection

Directory of Open Access Journals (Sweden)

Ralf C. Staudemeyer

2015-07-01

Full Text Available We claim that modelling network traffic as a time series with a supervised learning approach, using known genuine and malicious behaviour, improves intrusion detection. To substantiate this, we trained long short-term memory (LSTM recurrent neural networks with the training data provided by the DARPA / KDD Cup ’99 challenge. To identify suitable LSTM-RNN network parameters and structure we experimented with various network topologies. We found networks with four memory blocks containing two cells each offer a good compromise between computational cost and detection performance. We applied forget gates and shortcut connections respectively. A learning rate of 0.1 and up to 1,000 epochs showed good results. We tested the performance on all features and on extracted minimal feature sets respectively. We evaluated different feature sets for the detection of all attacks within one network and also to train networks specialised on individual attack classes. Our results show that the LSTM classifier provides superior performance in comparison to results previously published results of strong static classifiers. With 93.82% accuracy and 22.13 cost, LSTM outperforms the winning entries of the KDD Cup ’99 challenge by far. This is due to the fact that LSTM learns to look back in time and correlate consecutive connection records. For the first time ever, we have demonstrated the usefulness of LSTM networks to intrusion detection.

Long-term intensive gymnastic training induced changes in intra- and inter-network functional connectivity: an independent component analysis.

Science.gov (United States)

Huang, Huiyuan; Wang, Junjing; Seger, Carol; Lu, Min; Deng, Feng; Wu, Xiaoyan; He, Yuan; Niu, Chen; Wang, Jun; Huang, Ruiwang

2018-01-01

Long-term intensive gymnastic training can induce brain structural and functional reorganization. Previous studies have identified structural and functional network differences between world class gymnasts (WCGs) and non-athletes at the whole-brain level. However, it is still unclear how interactions within and between functional networks are affected by long-term intensive gymnastic training. We examined both intra- and inter-network functional connectivity of gymnasts relative to non-athletes using resting-state fMRI (R-fMRI). R-fMRI data were acquired from 13 WCGs and 14 non-athlete controls. Group-independent component analysis (ICA) was adopted to decompose the R-fMRI data into spatial independent components and associated time courses. An automatic component identification method was used to identify components of interest associated with resting-state networks (RSNs). We identified nine RSNs, the basal ganglia network (BG), sensorimotor network (SMN), cerebellum (CB), anterior and posterior default mode networks (aDMN/pDMN), left and right fronto-parietal networks (lFPN/rFPN), primary visual network (PVN), and extrastriate visual network (EVN). Statistical analyses revealed that the intra-network functional connectivity was significantly decreased within the BG, aDMN, lFPN, and rFPN, but increased within the EVN in the WCGs compared to the controls. In addition, the WCGs showed uniformly decreased inter-network functional connectivity between SMN and BG, CB, and PVN, BG and PVN, and pDMN and rFPN compared to the controls. We interpret this generally weaker intra- and inter-network functional connectivity in WCGs during the resting state as a result of greater efficiency in the WCGs' brain associated with long-term motor skill training.

Outcomes from the GLEON fellowship program. Training graduate students in data driven network science.

Science.gov (United States)

Dugan, H.; Hanson, P. C.; Weathers, K. C.

2016-12-01

In the water sciences there is a massive need for graduate students who possess the analytical and technical skills to deal with large datasets and function in the new paradigm of open, collaborative -science. The Global Lake Ecological Observatory Network (GLEON) graduate fellowship program (GFP) was developed as an interdisciplinary training program to supplement the intensive disciplinary training of traditional graduate education. The primary goal of the GFP was to train a diverse cohort of graduate students in network science, open-web technologies, collaboration, and data analytics, and importantly to provide the opportunity to use these skills to conduct collaborative research resulting in publishable scientific products. The GFP is run as a series of three week-long workshops over two years that brings together a cohort of twelve students. In addition, fellows are expected to attend and contribute to at least one international GLEON all-hands' meeting. Here, we provide examples of training modules in the GFP (model building, data QA/QC, information management, bayesian modeling, open coding/version control, national data programs), as well as scientific outputs (manuscripts, software products, and new global datasets) produced by the fellows, as well as the process by which this team science was catalyzed. Data driven education that lets students apply learned skills to real research projects reinforces concepts, provides motivation, and can benefit their publication record. This program design is extendable to other institutions and networks.

A stochastic learning algorithm for layered neural networks

International Nuclear Information System (INIS)

Bartlett, E.B.; Uhrig, R.E.

1992-01-01

The random optimization method typically uses a Gaussian probability density function (PDF) to generate a random search vector. In this paper the random search technique is applied to the neural network training problem and is modified to dynamically seek out the optimal probability density function (OPDF) from which to select the search vector. The dynamic OPDF search process, combined with an auto-adaptive stratified sampling technique and a dynamic node architecture (DNA) learning scheme, completes the modifications of the basic method. The DNA technique determines the appropriate number of hidden nodes needed for a given training problem. By using DNA, researchers do not have to set the neural network architectures before training is initiated. The approach is applied to networks of generalized, fully interconnected, continuous perceptions. Computer simulation results are given

Application of neural network technology to setpoint control of a simulated reactor experiment loop

International Nuclear Information System (INIS)

Cordes, G.A.; Bryan, S.R.; Powell, R.H.; Chick, D.R.

1991-01-01

This paper describes the design, implementation, and application of artificial neural networks to achieve temperature and flow rate control for a simulation of a typical experiment loop in the Advanced Test Reactor (ATR) located at the Idaho National Engineering Laboratory (INEL). The goal of the project was to research multivariate, nonlinear control using neural networks. A loop simulation code was adapted for the project and used to create a training set and test the neural network controller for comparison with the existing loop controllers. The results for the best neural network design are documented and compared with existing loop controller action. The neural network was shown to be as accurate at loop control as the classical controllers in the operating region represented by the training set. 5 refs., 8 figs., 3 tabs

Timetable-based simulation method for choice set generation in large-scale public transport networks

DEFF Research Database (Denmark)

Rasmussen, Thomas Kjær; Anderson, Marie Karen; Nielsen, Otto Anker

2016-01-01

The composition and size of the choice sets are a key for the correct estimation of and prediction by route choice models. While existing literature has posed a great deal of attention towards the generation of path choice sets for private transport problems, the same does not apply to public...... transport problems. This study proposes a timetable-based simulation method for generating path choice sets in a multimodal public transport network. Moreover, this study illustrates the feasibility of its implementation by applying the method to reproduce 5131 real-life trips in the Greater Copenhagen Area...... and to assess the choice set quality in a complex multimodal transport network. Results illustrate the applicability of the algorithm and the relevance of the utility specification chosen for the reproduction of real-life path choices. Moreover, results show that the level of stochasticity used in choice set...

Epidemiologic comparison of injured high school basketball athletes reporting to emergency departments and the athletic training setting.

Science.gov (United States)

Fletcher, Erica N; McKenzie, Lara B; Comstock, R Dawn

2014-01-01

Basketball is a popular US high school sport with more than 1 million participants annually. To compare patterns of athletes with basketball-related injuries presenting to US emergency departments from 2005 through 2010 and the high school athletic training setting from the 2005-2011 seasons. Descriptive epidemiology study. Data from the National Electronic Injury Surveillance System of the US Consumer Product Safety Commission and the High School Reporting Information Online database. Complex sample weights were used to calculate national estimates of basketball-related injuries for comparison. Adolescents from 13 to 19 years of age treated in US emergency departments for basketball-related injuries and athletes from 13 to 19 years of age from schools participating in High School Reporting Information Online who were injured while playing basketball. Nationally, an estimated 1,514,957 (95% confidence interval = 1,337,441, 1,692,474) athletes with basketball-related injuries reported to the emergency department and 1,064,551 (95% confidence interval = 1,055,482, 1,073,620) presented to the athletic training setting. Overall, the most frequent injuries seen in the emergency department were lacerations and fractures (injury proportion ratios [IPRs] = 3.45 and 1.72, respectively), whereas those seen in the athletic training setting were more commonly concussions and strains/sprains (IPRs = 2.23 and 1.19, respectively; all P values training setting (IPR = 1.18; all P values basketball players presenting for treatment in the emergency department and the athletic training setting. Understanding differences specific to clinical settings is crucial to grasping the full epidemiologic and clinical picture of sport-related injuries. Certified athletic trainers play an important role in identifying, assessing, and treating athletes with sport-related injuries who might otherwise present to clinical settings with higher costs, such as the emergency department.

Fast neutron spectra determination by threshold activation detectors using neural networks

International Nuclear Information System (INIS)

Kardan, M.R.; Koohi-Fayegh, R.; Setayeshi, S.; Ghiassi-Nejad, M.

2004-01-01

Neural network method was used for fast neutron spectra unfolding in spectrometry by threshold activation detectors. The input layer of the neural networks consisted of 11 neurons for the specific activities of neutron-induced nuclear reaction products, while the output layers were fast neutron spectra which had been subdivided into 6, 8, 10, 12, 15 and 20 energy bins. Neural network training was performed by 437 fast neutron spectra and corresponding threshold activation detector readings. The trained neural network have been applied for unfolding 50 spectra, which were not in training sets and the results were compared with real spectra and unfolded spectra by SANDII. The best results belong to 10 energy bin spectra. The neural network was also trained by detector readings with 5% uncertainty and the response of the trained neural network to detector readings with 5%, 10%, 15%, 20%, 25% and 50% uncertainty was compared with real spectra. Neural network algorithm, in comparison with other unfolding methods, is very fast and needless to detector response matrix and any prior information about spectra and also the outputs have low sensitivity to uncertainty in the activity measurements. The results show that the neural network algorithm is useful when a fast response is required with reasonable accuracy

A novel approach for voltage secure operation using Probabilistic Neural Network in transmission network

Directory of Open Access Journals (Sweden)

Santi Behera

2016-05-01

Full Text Available This work proposes a unique approach for improving voltage stability limit using a Probabilistic Neural Network (PNN classifier that gives corrective controls available in the system in the scenario of contingencies. The sensitivity of system is analyzed to identify weak buses with ENVCI evaluation approaching zero. The input to the classifier, termed as voltage stability enhancing neural network (VSENN classifier, for training are line flows and bus voltages near the notch point of the P–V curve and the output of the VSENN is a control variable. For various contingencies the control action that improves the voltage profile as well as stability index is identified and trained accordingly. The trained VSENN is finally tested for its robustness to improve load margin and ENVCI as well, apart from trained set of operating condition of the system along with contingencies. The proposed approach is verified in IEEE 39-bus test system.

Identification of Abnormal System Noise Temperature Patterns in Deep Space Network Antennas Using Neural Network Trained Fuzzy Logic

Science.gov (United States)

Lu, Thomas; Pham, Timothy; Liao, Jason

2011-01-01

This paper presents the development of a fuzzy logic function trained by an artificial neural network to classify the system noise temperature (SNT) of antennas in the NASA Deep Space Network (DSN). The SNT data were classified into normal, marginal, and abnormal classes. The irregular SNT pattern was further correlated with link margin and weather data. A reasonably good correlation is detected among high SNT, low link margin and the effect of bad weather; however we also saw some unexpected non-correlations which merit further study in the future.

Neural Network Emulation of Reionization Simulations

Science.gov (United States)

Schmit, Claude J.; Pritchard, Jonathan R.

2018-05-01

Next generation radio experiments such as LOFAR, HERA and SKA are expected to probe the Epoch of Reionization and claim a first direct detection of the cosmic 21cm signal within the next decade. One of the major challenges for these experiments will be dealing with enormous incoming data volumes. Machine learning is key to increasing our data analysis efficiency. We consider the use of an artificial neural network to emulate 21cmFAST simulations and use it in a Bayesian parameter inference study. We then compare the network predictions to a direct evaluation of the EoR simulations and analyse the dependence of the results on the training set size. We find that the use of a training set of size 100 samples can recover the error contours of a full scale MCMC analysis which evaluates the model at each step.

Wind power forecast using wavelet neural network trained by improved Clonal selection algorithm

International Nuclear Information System (INIS)

Chitsaz, Hamed; Amjady, Nima; Zareipour, Hamidreza

2015-01-01

Highlights: • Presenting a Morlet wavelet neural network for wind power forecasting. • Proposing improved Clonal selection algorithm for training the model. • Applying Maximum Correntropy Criterion to evaluate the training performance. • Extensive testing of the proposed wind power forecast method on real-world data. - Abstract: With the integration of wind farms into electric power grids, an accurate wind power prediction is becoming increasingly important for the operation of these power plants. In this paper, a new forecasting engine for wind power prediction is proposed. The proposed engine has the structure of Wavelet Neural Network (WNN) with the activation functions of the hidden neurons constructed based on multi-dimensional Morlet wavelets. This forecast engine is trained by a new improved Clonal selection algorithm, which optimizes the free parameters of the WNN for wind power prediction. Furthermore, Maximum Correntropy Criterion (MCC) has been utilized instead of Mean Squared Error as the error measure in training phase of the forecasting model. The proposed wind power forecaster is tested with real-world hourly data of system level wind power generation in Alberta, Canada. In order to demonstrate the efficiency of the proposed method, it is compared with several other wind power forecast techniques. The obtained results confirm the validity of the developed approach

Monitoring of Students' Interaction in Online Learning Settings by Structural Network Analysis and Indicators.

Science.gov (United States)

Ammenwerth, Elske; Hackl, Werner O

2017-01-01

Learning as a constructive process works best in interaction with other learners. Support of social interaction processes is a particular challenge within online learning settings due to the spatial and temporal distribution of participants. It should thus be carefully monitored. We present structural network analysis and related indicators to analyse and visualize interaction patterns of participants in online learning settings. We validate this approach in two online courses and show how the visualization helps to monitor interaction and to identify activity profiles of learners. Structural network analysis is a feasible approach for an analysis of the intensity and direction of interaction in online learning settings.

Enhancing adaptive capacity for restoring fire-dependent ecosystems: the Fire Learning Network's Prescribed Fire Training Exchanges

Directory of Open Access Journals (Sweden)

Andrew G. Spencer

2015-09-01

Full Text Available Prescribed fire is a critical tool for promoting restoration and increasing resilience in fire-adapted ecosystems, but there are barriers to its use, including a shortage of personnel with adequate ecological knowledge and operational expertise to implement prescribed fire across multijurisdictional landscapes. In the United States, recognized needs for both professional development and increased use of fire are not being met, often because of institutional limitations. The Fire Learning Network has been characterized as a multiscalar, collaborative network that works to enhance the adaptive capacity of fire management institutions, and this network developed the Prescribed Fire Training Exchanges (TREXs to address persistent challenges in increasing the capacity for prescribed fire implementation. Our research was designed to investigate where fire professionals face professional barriers, how the TREX addresses these, and in what ways the TREX may be contributing to the adaptive capacity of fire management institutions. We evaluated the training model using surveys, interviews, focus groups, and participant observation. We found that, although the training events cannot overcome all institutional barriers, they incorporate the key components of professional development in fire; foster collaboration, learning, and network building; and provide flexible opportunities with an emphasis on local context to train a variety of professionals with disparate needs. The strategy also offers an avenue for overcoming barriers faced by contingent and nonfederal fire professionals in attaining training and operational experience, thereby increasing the variety of actors and resources involved in fire management. Although it is an incremental step, the TREX is contributing to the adaptive capacity of institutions in social-ecological systems in which fire is a critical ecological process.

A microsensor array for quantification of lubricant contaminants using a back propagation artificial neural network

International Nuclear Information System (INIS)

Zhu, Xiaoliang; Du, Li; Zhe, Jiang; Liu, Bendong

2016-01-01

We present a method based on an electrochemical sensor array and a back propagation artificial neural network for detection and quantification of four properties of lubrication oil, namely water (0, 500 ppm, 1000 ppm), total acid number (TAN) (13.1, 13.7, 14.4, 15.6 mg KOH g −1 ), soot (0, 1%, 2%, 3%) and sulfur content (1.3%, 1.37%, 1.44%, 1.51%). The sensor array, consisting of four micromachined electrochemical sensors, detects the four properties with overlapping sensitivities. A total set of 36 oil samples containing mixtures of water, soot, and sulfuric acid with different concentrations were prepared for testing. The sensor array’s responses were then divided to three sets: training sets (80% data), validation sets (10%) and testing sets (10%). Several back propagation artificial neural network architectures were trained with the training and validation sets; one architecture with four input neurons, 50 and 5 neurons in the first and second hidden layer, and four neurons in the output layer was selected. The selected neural network was then tested using the four sets of testing data (10%). Test results demonstrated that the developed artificial neural network is able to quantitatively determine the four lubrication properties (water, TAN, soot, and sulfur content) with a maximum prediction error of 18.8%, 6.0%, 6.7%, and 5.4%, respectively, indicting a good match between the target and predicted values. With the developed network, the sensor array could be potentially used for online lubricant oil condition monitoring. (paper)

Let’s Wiggle with 5-2-1-0: Curriculum Development for Training Childcare Providers to Promote Activity in Childcare Settings

Directory of Open Access Journals (Sweden)

Debra M. Vinci

2016-01-01

Full Text Available Overweight and obesity are increasing in preschool children in the US. Policy, systems, and environmental change interventions in childcare settings can improve obesity-related behaviors. The aim of this study was to develop and pilot an intervention to train childcare providers to promote physical activity (PA in childcare classrooms. An evidence scan, key informant (n=34 and focus group (n=20 interviews with childcare directors and staff, and environmental self-assessment of childcare facilities (n=22 informed the design of the training curriculum. Feedback from the interviews indicated that childcare providers believed in the importance of teaching children about PA and were supportive of training teachers to incorporate PA into classroom settings. The Promoting Physical Activity in Childcare Setting Curriculum was developed and training was implemented with 16 teachers. Participants reported a positive experience with the hands-on training and reported acquiring new knowledge that they intended to implement in their childcare settings. Our findings highlight the feasibility of working with childcare staff to develop PA training and curriculum. Next steps include evaluating the curriculum in additional childcare settings and childcare staff implementation of the curriculum to understand the effectiveness of the training on PA levels of children.

Pre-trained convolutional neural networks as feature extractors for tuberculosis detection.

Science.gov (United States)

Lopes, U K; Valiati, J F

2017-10-01

It is estimated that in 2015, approximately 1.8 million people infected by tuberculosis died, most of them in developing countries. Many of those deaths could have been prevented if the disease had been detected at an earlier stage, but the most advanced diagnosis methods are still cost prohibitive for mass adoption. One of the most popular tuberculosis diagnosis methods is the analysis of frontal thoracic radiographs; however, the impact of this method is diminished by the need for individual analysis of each radiography by properly trained radiologists. Significant research can be found on automating diagnosis by applying computational techniques to medical images, thereby eliminating the need for individual image analysis and greatly diminishing overall costs. In addition, recent improvements on deep learning accomplished excellent results classifying images on diverse domains, but its application for tuberculosis diagnosis remains limited. Thus, the focus of this work is to produce an investigation that will advance the research in the area, presenting three proposals to the application of pre-trained convolutional neural networks as feature extractors to detect the disease. The proposals presented in this work are implemented and compared to the current literature. The obtained results are competitive with published works demonstrating the potential of pre-trained convolutional networks as medical image feature extractors. Copyright © 2017 Elsevier Ltd. All rights reserved.

Enlarge the training set based on inter-class relationship for face recognition from one image per person.

Science.gov (United States)

Li, Qin; Wang, Hua Jing; You, Jane; Li, Zhao Ming; Li, Jin Xue

2013-01-01

In some large-scale face recognition task, such as driver license identification and law enforcement, the training set only contains one image per person. This situation is referred to as one sample problem. Because many face recognition techniques implicitly assume that several (at least two) images per person are available for training, they cannot deal with the one sample problem. This paper investigates principal component analysis (PCA), Fisher linear discriminant analysis (LDA), and locality preserving projections (LPP) and shows why they cannot perform well in one sample problem. After that, this paper presents four reasons that make one sample problem itself difficult: the small sample size problem; the lack of representative samples; the underestimated intra-class variation; and the overestimated inter-class variation. Based on the analysis, this paper proposes to enlarge the training set based on the inter-class relationship. This paper also extends LDA and LPP to extract features from the enlarged training set. The experimental results show the effectiveness of the proposed method.

Efficient batchwise dropout training using submatrices

OpenAIRE

Graham, Ben; Reizenstein, Jeremy; Robinson, Leigh

2015-01-01

Dropout is a popular technique for regularizing artificial neural networks. Dropout networks are generally trained by minibatch gradient descent with a dropout mask turning off some of the units---a different pattern of dropout is applied to every sample in the minibatch. We explore a very simple alternative to the dropout mask. Instead of masking dropped out units by setting them to zero, we perform matrix multiplication using a submatrix of the weight matrix---unneeded hidden units are neve...

Physicists set new record for network data transfer

CERN Multimedia

2006-01-01

"An internatinal team of physicists, computer scientists, and network engineers led by the California Institute of Technology, CERN and the University of Michigan and partners at the University of Florida and Vanderbilt, as well as participants from Brazil (Rio de Janeiro State University, UERJ, and the State Universities of Sao Paulo, USP and UNESP) and Korea (Kyungpook National University, KISTI) joined forces to set new records for sustained data transfer between storage systems during the SuperComputing 2006 (SC06) Bandwidth Challenge (BWC)." (2 pages)

International Nuclear Security Education Network (INSEN) and the Nuclear Security Training and Support Centre (NSSC) Network

International Nuclear Information System (INIS)

Nikonov, Dmitriy

2013-01-01

International Nuclear Security Education Network established in 2010: A partnership between the IAEA and universities, research institutions and other stakeholders - •Promotion of nuclear security education; • Development of educational materials; • Professional development for faculty members; • Collaborative research and resource sharing. Currently over 90 members from 38 member states. Mission: to enhance global nuclear security by developing, sharing and promoting excellence in nuclear security education. Nuclear Security Support Centre: Primary objectives are: • Develop human resources through the implementation of a tailored training programme; • Develop a network of experts; • Provide technical support for lifecycle equipment management and scientific support for the detection of and the response to nuclear security events

Evolutionary Algorithms For Neural Networks Binary And Real Data Classification

Directory of Open Access Journals (Sweden)

Dr. Hanan A.R. Akkar

2015-08-01

Full Text Available Artificial neural networks are complex networks emulating the way human rational neurons process data. They have been widely used generally in prediction clustering classification and association. The training algorithms that used to determine the network weights are almost the most important factor that influence the neural networks performance. Recently many meta-heuristic and Evolutionary algorithms are employed to optimize neural networks weights to achieve better neural performance. This paper aims to use recently proposed algorithms for optimizing neural networks weights comparing these algorithms performance with other classical meta-heuristic algorithms used for the same purpose. However to evaluate the performance of such algorithms for training neural networks we examine such algorithms to classify four opposite binary XOR clusters and classification of continuous real data sets such as Iris and Ecoli.

Towards dropout training for convolutional neural networks.

Science.gov (United States)

Wu, Haibing; Gu, Xiaodong

2015-11-01

Recently, dropout has seen increasing use in deep learning. For deep convolutional neural networks, dropout is known to work well in fully-connected layers. However, its effect in convolutional and pooling layers is still not clear. This paper demonstrates that max-pooling dropout is equivalent to randomly picking activation based on a multinomial distribution at training time. In light of this insight, we advocate employing our proposed probabilistic weighted pooling, instead of commonly used max-pooling, to act as model averaging at test time. Empirical evidence validates the superiority of probabilistic weighted pooling. We also empirically show that the effect of convolutional dropout is not trivial, despite the dramatically reduced possibility of over-fitting due to the convolutional architecture. Elaborately designing dropout training simultaneously in max-pooling and fully-connected layers, we achieve state-of-the-art performance on MNIST, and very competitive results on CIFAR-10 and CIFAR-100, relative to other approaches without data augmentation. Finally, we compare max-pooling dropout and stochastic pooling, both of which introduce stochasticity based on multinomial distributions at pooling stage. Copyright © 2015 Elsevier Ltd. All rights reserved.

Determination of fluence-to-dose conversion coefficients by means of artificial neural networks

International Nuclear Information System (INIS)

Soto B, T. G.; Rivera P, E.; De Leon M, H. A.; Hernandez D, V. M.; Vega C, H. R.; Gallego, E.; Lorente, A.

2012-10-01

In this paper is presented an Artificial Neural Network (Ann) that has been designed, trained and validated to determinate the effective dose e, ambient dose equivalent h(10) and personal dose equivalent hp(10,θ) fluence-to-dose conversion coefficients at different positions, having as only input data 7 count rates obtained with a Bonner Sphere Spectrometer (Bss) system. A set of 211 neutron spectra and the fluence-to-dose conversion coefficients published by the International Atomic Energy Agency were used to train and validate the Ann. This set was divided into 2 subsets, one of 181 elements to train the Ann and the remaining 30 to validate it. The Ann was trained using Bss count rates as input data and the fluence-to-dose conversion coefficients as output data. The network was validated and tested with the set of 30 elements that were not used during the training process. Good results were obtained proving that Ann are a good choice for calculating the fluence-to-dose conversion coefficients having as only data the count rates obtained with a Bss. (Author)

Determination of fluence-to-dose conversion coefficients by means of artificial neural networks

Energy Technology Data Exchange (ETDEWEB)

Soto B, T. G.; Rivera P, E.; De Leon M, H. A.; Hernandez D, V. M.; Vega C, H. R. [Universidad Autonoma de Zacatecas, Unidad Academica de Estudios Nucleares, Cipres No. 10, Fracc. La Penuela, 98068 Zacatecas (Mexico); Gallego, E.; Lorente, A., E-mail: tzinnia.soto@gmail.com [Universidad Politecnica de Madrid, Departamento de Ingenieria Nuclear, Jose Gutierrez Abascal No. 2, 28006 Madrid (Spain)

2012-10-15

In this paper is presented an Artificial Neural Network (Ann) that has been designed, trained and validated to determinate the effective dose e, ambient dose equivalent h(10) and personal dose equivalent hp(10,{theta}) fluence-to-dose conversion coefficients at different positions, having as only input data 7 count rates obtained with a Bonner Sphere Spectrometer (Bss) system. A set of 211 neutron spectra and the fluence-to-dose conversion coefficients published by the International Atomic Energy Agency were used to train and validate the Ann. This set was divided into 2 subsets, one of 181 elements to train the Ann and the remaining 30 to validate it. The Ann was trained using Bss count rates as input data and the fluence-to-dose conversion coefficients as output data. The network was validated and tested with the set of 30 elements that were not used during the training process. Good results were obtained proving that Ann are a good choice for calculating the fluence-to-dose conversion coefficients having as only data the count rates obtained with a Bss. (Author)

A managed clinical network for cardiac services: set-up, operation and impact on patient care

Directory of Open Access Journals (Sweden)

Karen E. Hamilton

2005-09-01

Full Text Available Purpose: To investigate the set up and operation of a Managed Clinical Network for cardiac services and assess its impact on patient care. Methods: This single case study used process evaluation with observational before and after comparison of indicators of quality of care and costs. The study was conducted in Dumfries and Galloway, Scotland and used a three-level framework. Process evaluation of the network set-up and operation through a documentary review of minutes; guidelines and protocols; transcripts of fourteen semi-structured interviews with health service personnel including senior managers, general practitioners, nurses, cardiologists and members of the public. Outcome evaluation of the impact of the network through interrupted time series analysis of clinical data of 202 patients aged less than 76 years admitted to hospital with a confirmed myocardial infarction one-year pre and one-year post, the establishment of the network. The main outcome measures were differences between indicators of quality of care targeted by network protocols. Economic evaluation of the transaction costs of the set-up and operation of the network and the resource costs of the clinical care of the 202 myocardial infarction patients from the time of hospital admission to 6 months post discharge through interrupted time series analysis. The outcome measure was different in National Health Service resource use. Results: Despite early difficulties, the network was successful in bringing together clinicians, patients and managers to redesign services, exhibiting most features of good network management. The role of the energetic lead clinician was crucial, but the network took time to develop and ‘bed down’. Its primary “modus operand” was the development of a myocardial infarction pathway and associated protocols. Of sixteen clinical care indicators, two improved significantly following the launch of the network and nine showed improvements, which were

Illusory Motion Reproduced by Deep Neural Networks Trained for Prediction.

Science.gov (United States)

Watanabe, Eiji; Kitaoka, Akiyoshi; Sakamoto, Kiwako; Yasugi, Masaki; Tanaka, Kenta

2018-01-01

The cerebral cortex predicts visual motion to adapt human behavior to surrounding objects moving in real time. Although the underlying mechanisms are still unknown, predictive coding is one of the leading theories. Predictive coding assumes that the brain's internal models (which are acquired through learning) predict the visual world at all times and that errors between the prediction and the actual sensory input further refine the internal models. In the past year, deep neural networks based on predictive coding were reported for a video prediction machine called PredNet. If the theory substantially reproduces the visual information processing of the cerebral cortex, then PredNet can be expected to represent the human visual perception of motion. In this study, PredNet was trained with natural scene videos of the self-motion of the viewer, and the motion prediction ability of the obtained computer model was verified using unlearned videos. We found that the computer model accurately predicted the magnitude and direction of motion of a rotating propeller in unlearned videos. Surprisingly, it also represented the rotational motion for illusion images that were not moving physically, much like human visual perception. While the trained network accurately reproduced the direction of illusory rotation, it did not detect motion components in negative control pictures wherein people do not perceive illusory motion. This research supports the exciting idea that the mechanism assumed by the predictive coding theory is one of basis of motion illusion generation. Using sensory illusions as indicators of human perception, deep neural networks are expected to contribute significantly to the development of brain research.

Associating Human-Centered Concepts with Social Networks Using Fuzzy Sets

Science.gov (United States)

Yager, Ronald R.

that allows us to determine how true it is that a particular node is a leader. In this work we look at the use of fuzzy set methodologies [8-10] to provide a bridge between the human analyst and the formal model of the network.

Cross-domain and multi-task transfer learning of deep convolutional neural network for breast cancer diagnosis in digital breast tomosynthesis

Science.gov (United States)

Samala, Ravi K.; Chan, Heang-Ping; Hadjiiski, Lubomir; Helvie, Mark A.; Richter, Caleb; Cha, Kenny

2018-02-01

We propose a cross-domain, multi-task transfer learning framework to transfer knowledge learned from non-medical images by a deep convolutional neural network (DCNN) to medical image recognition task while improving the generalization by multi-task learning of auxiliary tasks. A first stage cross-domain transfer learning was initiated from ImageNet trained DCNN to mammography trained DCNN. 19,632 regions-of-interest (ROI) from 2,454 mass lesions were collected from two imaging modalities: digitized-screen film mammography (SFM) and full-field digital mammography (DM), and split into training and test sets. In the multi-task transfer learning, the DCNN learned the mass classification task simultaneously from the training set of SFM and DM. The best transfer network for mammography was selected from three transfer networks with different number of convolutional layers frozen. The performance of single-task and multitask transfer learning on an independent SFM test set in terms of the area under the receiver operating characteristic curve (AUC) was 0.78+/-0.02 and 0.82+/-0.02, respectively. In the second stage cross-domain transfer learning, a set of 12,680 ROIs from 317 mass lesions on DBT were split into validation and independent test sets. We first studied the data requirements for the first stage mammography trained DCNN by varying the mammography training data from 1% to 100% and evaluated its learning on the DBT validation set in inference mode. We found that the entire available mammography set provided the best generalization. The DBT validation set was then used to train only the last four fully connected layers, resulting in an AUC of 0.90+/-0.04 on the independent DBT test set.

Weighted complex network analysis of the Beijing subway system: Train and passenger flows

Science.gov (United States)

Feng, Jia; Li, Xiamiao; Mao, Baohua; Xu, Qi; Bai, Yun

2017-05-01

In recent years, complex network theory has become an important approach to the study of the structure and dynamics of traffic networks. However, because traffic data is difficult to collect, previous studies have usually focused on the physical topology of subway systems, whereas few studies have considered the characteristics of traffic flows through the network. Therefore, in this paper, we present a multi-layer model to analyze traffic flow patterns in subway networks, based on trip data and an operation timetable obtained from the Beijing Subway System. We characterize the patterns in terms of the spatiotemporal flow size distributions of both the train flow network and the passenger flow network. In addition, we describe the essential interactions between these two networks based on statistical analyses. The results of this study suggest that layered models of transportation systems can elucidate fundamental differences between the coexisting traffic flows and can also clarify the mechanism that causes these differences.

Improving quantitative structure-activity relationship models using Artificial Neural Networks trained with dropout.

Science.gov (United States)

Mendenhall, Jeffrey; Meiler, Jens

2016-02-01

Dropout is an Artificial Neural Network (ANN) training technique that has been shown to improve ANN performance across canonical machine learning (ML) datasets. Quantitative Structure Activity Relationship (QSAR) datasets used to relate chemical structure to biological activity in Ligand-Based Computer-Aided Drug Discovery pose unique challenges for ML techniques, such as heavily biased dataset composition, and relatively large number of descriptors relative to the number of actives. To test the hypothesis that dropout also improves QSAR ANNs, we conduct a benchmark on nine large QSAR datasets. Use of dropout improved both enrichment false positive rate and log-scaled area under the receiver-operating characteristic curve (logAUC) by 22-46 % over conventional ANN implementations. Optimal dropout rates are found to be a function of the signal-to-noise ratio of the descriptor set, and relatively independent of the dataset. Dropout ANNs with 2D and 3D autocorrelation descriptors outperform conventional ANNs as well as optimized fingerprint similarity search methods.

Massive training artificial neural network (MTANN) for reduction of false positives in computerized detection of lung nodules in low-dose computed tomography

International Nuclear Information System (INIS)

Suzuki, Kenji; Armato, Samuel G. III; Li, Feng; Sone, Shusuke; Doi, Kunio

2003-01-01

In this study, we investigated a pattern-recognition technique based on an artificial neural network (ANN), which is called a massive training artificial neural network (MTANN), for reduction of false positives in computerized detection of lung nodules in low-dose computed tomography (CT) images. The MTANN consists of a modified multilayer ANN, which is capable of operating on image data directly. The MTANN is trained by use of a large number of subregions extracted from input images together with the teacher images containing the distribution for the 'likelihood of being a nodule'. The output image is obtained by scanning an input image with the MTANN. The distinction between a nodule and a non-nodule is made by use of a score which is defined from the output image of the trained MTANN. In order to eliminate various types of non-nodules, we extended the capability of a single MTANN, and developed a multiple MTANN (Multi-MTANN). The Multi-MTANN consists of plural MTANNs that are arranged in parallel. Each MTANN is trained by using the same nodules, but with a different type of non-nodule. Each MTANN acts as an expert for a specific type of non-nodule, e.g., five different MTANNs were trained to distinguish nodules from various-sized vessels; four other MTANNs were applied to eliminate some other opacities. The outputs of the MTANNs were combined by using the logical AND operation such that each of the trained MTANNs eliminated none of the nodules, but removed the specific type of non-nodule with which the MTANN was trained, and thus removed various types of non-nodules. The Multi-MTANN consisting of nine MTANNs was trained with 10 typical nodules and 10 non-nodules representing each of nine different non-nodule types (90 training non-nodules overall) in a training set. The trained Multi-MTANN was applied to the reduction of false positives reported by our current computerized scheme for lung nodule detection based on a database of 63 low-dose CT scans (1765

Cascading Generative Adversarial Networks for Targeted

KAUST Repository

Hamdi, Abdullah

2018-01-01

Abundance of labelled data played a crucial role in the recent developments in computer vision, but that faces problems like scalability and transferability to the wild. One alternative approach is to utilize the data without labels, i.e. unsupervised learning, in learning valuable information and put it in use to tackle vision problems. Generative Adversarial Networks (GANs) have gained momentum for their ability to model image distributions in unsupervised manner. They learn to emulate the training set and that enables sampling from that domain and using the knowledge learned for useful applications. Several methods proposed enhancing GANs, including regularizing the loss with some feature matching. We seek to push GANs beyond the data in the training and try to explore unseen territory in the image manifold. We first propose a new regularizer for GAN based on K-Nearest Neighbor (K-NN) selective feature matching to a target set Y in high-level feature space, during the adversarial training of GAN on the base set X, and we call this novel model K-GAN. We show that minimizing the added term follows from cross-entropy minimization between the distributions of GAN and set Y. Then, we introduce a cascaded framework for GANs that try to address the task of imagining a new distribution that combines the base set X and target set Y by cascading sampling GANs with translation GANs, and we dub the cascade of such GANs as the Imaginative Adversarial Network (IAN). Several cascades are trained on a collected dataset Zoo-Faces and generated innovative samples are shown, including from K-GAN cascade. We conduct an objective and subjective evaluation for different IAN setups in the addressed task of generating innovative samples and we show the effect of regularizing GAN on different scores. We conclude with some useful applications for these IANs, like multi-domain manifold traversing.

Cascading Generative Adversarial Networks for Targeted

KAUST Repository

Hamdi, Abdullah

2018-04-09

Abundance of labelled data played a crucial role in the recent developments in computer vision, but that faces problems like scalability and transferability to the wild. One alternative approach is to utilize the data without labels, i.e. unsupervised learning, in learning valuable information and put it in use to tackle vision problems. Generative Adversarial Networks (GANs) have gained momentum for their ability to model image distributions in unsupervised manner. They learn to emulate the training set and that enables sampling from that domain and using the knowledge learned for useful applications. Several methods proposed enhancing GANs, including regularizing the loss with some feature matching. We seek to push GANs beyond the data in the training and try to explore unseen territory in the image manifold. We first propose a new regularizer for GAN based on K-Nearest Neighbor (K-NN) selective feature matching to a target set Y in high-level feature space, during the adversarial training of GAN on the base set X, and we call this novel model K-GAN. We show that minimizing the added term follows from cross-entropy minimization between the distributions of GAN and set Y. Then, we introduce a cascaded framework for GANs that try to address the task of imagining a new distribution that combines the base set X and target set Y by cascading sampling GANs with translation GANs, and we dub the cascade of such GANs as the Imaginative Adversarial Network (IAN). Several cascades are trained on a collected dataset Zoo-Faces and generated innovative samples are shown, including from K-GAN cascade. We conduct an objective and subjective evaluation for different IAN setups in the addressed task of generating innovative samples and we show the effect of regularizing GAN on different scores. We conclude with some useful applications for these IANs, like multi-domain manifold traversing.

Evaluating Question, Persuade, Refer (QPR) Suicide Prevention Training in a College Setting

Science.gov (United States)

Mitchell, Sharon L.; Kader, Mahrin; Darrow, Sherri A.; Haggerty, Melinda Z.; Keating, Niki L.

2013-01-01

This study assesses short-term and long-term learning outcomes of Question, Persuade, Refer (QPR) suicide prevention training in a college setting. Two hundred seventy-three participants completed pretest, posttest, and follow-up surveys regarding suicide prevention knowledge, attitudes, and skills. Results indicated: (a) increases in suicide…

Efficient Gatherings in Wireless Sensor Networks Using Distributed Computation of Connected Dominating Sets

Directory of Open Access Journals (Sweden)

Vincent BOUDET

2012-03-01

Full Text Available In this paper, we are interested in enhancing lifetime of wireless sensor networks trying to collect data from all the nodes to a “sink”-node for non-safety critical applications. Connected Dominating Sets are used as a basis for routing messages to the sink. We present a simple distributed algorithm, which computes several CDS trying to distribute the consumption of energy over all the nodes of the network. The simulations show a significant improvement of the network lifetime.

Multimedia Technologies for Training: An Introduction.

Science.gov (United States)

Barron, Ann E.; Orwig, Gary W.

This guide introduces trainers, managers, and educators to a variety of new multimedia technologies now being used for presentation and training in business, military, and academic settings. The text describes advances in and implementation of technologies that range from wireless local area networks (LANs) and high definition television (HDTV) to…

Navigating Social Networking and Social Media in School Psychology: Ethical and Professional Considerations in Training Programs

Science.gov (United States)

Pham, Andy V.

2014-01-01

Social networking and social media have undoubtedly proliferated within the past decade, allowing widespread communication and dissemination of user-generated content and information. Some psychology graduate programs, including school psychology, have started to embrace social networking and media for instructional and training purposes; however,…

Tuning of spinal networks to frequency components of spike trains in individual afferents.

Science.gov (United States)

Koerber, H R; Seymour, A W; Mendell, L M

1991-10-01

Cord dorsum potentials (CDPs) evoked by primary afferent fiber stimulation reflect the response of postsynaptic dorsal horn neurons. The properties of these CDPs have been shown to vary in accordance with the type of primary afferent fiber stimulated. The purpose of the present study was to determine the relationships between frequency modulation of the afferent input trains, the amplitude modulation of the evoked CDPs, and the type of primary afferent stimulated. The somata of individual primary afferent fibers were impaled in the L7 dorsal root ganglion of alpha-chloralose-anesthetized cats. Action potentials (APs) were evoked in single identified afferents via the intracellular microelectrode while simultaneously recording the response of dorsal horn neurons as CDPs, or activity of individual target interneurons recorded extracellularly or intracellularly. APs were evoked in afferents using temporal patterns identical to the responses of selected afferents to natural stimulation of their receptive fields. Two such physiologically realistic trains, one recorded from a hair follicle and the other from a slowly adapting type 1 receptor, were chosen as standard test trains. Modulation of CDP amplitude in response to this frequency-modulated afferent activity varied according to the type of peripheral mechanoreceptor innervated. Dorsal horn networks driven by A beta afferents innervating hair follicles, rapidly adapting pad (Krause end bulb), and field receptors seemed "tuned" to amplify the onset of activity in single afferents. Networks driven by afferents innervating down hair follicles and pacinian corpuscles required more high-frequency activity to elicit their peak response. Dorsal horn networks driven by afferents innervating slowly adapting receptors including high-threshold mechanoreceptors exhibited some sensitivity to the instantaneous frequency, but in general they reproduced the activity in the afferent fiber much more faithfully. Responses of

A Hybrid Spectral Clustering and Deep Neural Network Ensemble Algorithm for Intrusion Detection in Sensor Networks.

Science.gov (United States)

Ma, Tao; Wang, Fen; Cheng, Jianjun; Yu, Yang; Chen, Xiaoyun

2016-10-13

The development of intrusion detection systems (IDS) that are adapted to allow routers and network defence systems to detect malicious network traffic disguised as network protocols or normal access is a critical challenge. This paper proposes a novel approach called SCDNN, which combines spectral clustering (SC) and deep neural network (DNN) algorithms. First, the dataset is divided into k subsets based on sample similarity using cluster centres, as in SC. Next, the distance between data points in a testing set and the training set is measured based on similarity features and is fed into the deep neural network algorithm for intrusion detection. Six KDD-Cup99 and NSL-KDD datasets and a sensor network dataset were employed to test the performance of the model. These experimental results indicate that the SCDNN classifier not only performs better than backpropagation neural network (BPNN), support vector machine (SVM), random forest (RF) and Bayes tree models in detection accuracy and the types of abnormal attacks found. It also provides an effective tool of study and analysis of intrusion detection in large networks.

The Bilevel Design Problem for Communication Networks on Trains: Model, Algorithm, and Verification

Directory of Open Access Journals (Sweden)

Yin Tian

2014-01-01

Full Text Available This paper proposes a novel method to solve the problem of train communication network design. Firstly, we put forward a general description of such problem. Then, taking advantage of the bilevel programming theory, we created the cost-reliability-delay model (CRD model that consisted of two parts: the physical topology part aimed at obtaining the networks with the maximum reliability under constrained cost, while the logical topology part focused on the communication paths yielding minimum delay based on the physical topology delivered from upper level. We also suggested a method to solve the CRD model, which combined the genetic algorithm and the Floyd-Warshall algorithm. Finally, we used a practical example to verify the accuracy and the effectiveness of the CRD model and further applied the novel method on a train with six carriages.

SET UP OF THE NEW AUTOMATIC HYDROMETEOROLOGICAL NETWORK IN HUNGARY

Directory of Open Access Journals (Sweden)

J. NAGy

2013-03-01

Full Text Available The Hungarian Meteorological Service (OMSZ and General Directorate of Water Management (OVF in Hungary run conventional precipitation measurement networks consisting of at least 1000 stations. OMSZ automated its synoptic and climatological network in 90’s and now more than 100 automatic stations give data every 1-10 minutes via GPRS channel. In 2007 the experts from both institutions determined the requirements of a common network. The predecessor in title of OVF is general Directorate for Water and Environment gave a project proposal in 2008 for establishment of a new hydrometeorological network based on common aims for meteorology and hydrology. The new hydrometeorological network was set up in 2012 financed by KEOP project. This network has got 141 weighing precipitation gauges, 118 temperature - humidity sensors and 25 soil moisture and soil temperature instruments. Near by Tisza-Lake two wind sensors have been installed. The network is operated by OMSZ and OVF together. OVF and its institutions maintain the stations itself and support the electricity. OMSZ operates data collection and transmission, maintaines and calibrates the sensors. Using precipitation data of enhanced network the radar precipitation field quality may be more precise, which are input of run-off model. Thereby the time allowance may be increased in flood-control events. Based on soil moisture and temperature water balance in soil may be modelled and forecast can be produced in different conditions. It is very important task in drought and inland water conditions. Considering OMSZ investment project in which new Doppler dual polarisation radar and 14 disdrometers will be installed, the precipitation estimation may be improved since 2015.

Automated contour detection in cardiac MRI using active appearance models: the effect of the composition of the training set

NARCIS (Netherlands)

Angelié, Emmanuelle; Oost, Elco R.; Hendriksen, Dennis; Lelieveldt, Boudewijn P. F.; van der Geest, Rob J.; Reiber, Johan H. C.

2007-01-01

Definition of the optimal training set for the automated segmentation of short-axis left ventricular magnetic resonance (MR) imaging studies in clinical practice based on active appearance model (AAM). We investigated the segmentation accuracy by varying the size and composition of the training set

Standardized network order sets in rural Ontario: a follow-up report on successes and sustainability.

Science.gov (United States)

Rawn, Andrea; Wilson, Katrina

2011-01-01

Unifying, implementing and sustaining a large order set project requires strategic placement of key organizational professionals to provide ongoing user education, communication and support. This article will outline the successful strategies implemented by the Grey Bruce Health Network, Evidence-Based Care Program to reduce length of stay, increase patient satisfaction and increase the use of best practices resulting in quality outcomes, safer practice and better allocation of resources by using standardized Order Sets within a network of 11 hospital sites. Audits conducted in 2007 and again in 2008 revealed a reduced length of stay of 0.96 in-patient days when order sets were used on admission and readmission for the same or a related diagnosis within one month decreased from 5.5% without order sets to 3.5% with order sets.

Composability-Centered Convolutional Neural Network Pruning

Energy Technology Data Exchange (ETDEWEB)

Shen, Xipeng [North Carolina State University; Guan, Hui [North Carolina State University; Lim, Seung-Hwan [ORNL; Patton, Robert M. [ORNL

2018-02-01

This work studies the composability of the building blocks ofstructural CNN models (e.g., GoogleLeNet and Residual Networks) in thecontext of network pruning. We empirically validate that a networkcomposed of pre-trained building blocks (e.g. residual blocks andInception modules) not only gives a better initial setting fortraining, but also allows the training process to converge at asignificantly higher accuracy in much less time. Based on thatinsight, we propose a {\\em composability-centered} design for CNNnetwork pruning. Experiments show that this new scheme shortens theconfiguration process in CNN network pruning by up to 186.8X forResNet-50 and up to 30.2X for Inception-V3, and meanwhile, the modelsit finds that meet the accuracy requirement are significantly morecompact than those found by default schemes.

Training Small Networks for Scene Classification of Remote Sensing Images via Knowledge Distillation

Directory of Open Access Journals (Sweden)

Guanzhou Chen

2018-05-01

Full Text Available Scene classification, aiming to identify the land-cover categories of remotely sensed image patches, is now a fundamental task in the remote sensing image analysis field. Deep-learning-model-based algorithms are widely applied in scene classification and achieve remarkable performance, but these high-level methods are computationally expensive and time-consuming. Consequently in this paper, we introduce a knowledge distillation framework, currently a mainstream model compression method, into remote sensing scene classification to improve the performance of smaller and shallower network models. Our knowledge distillation training method makes the high-temperature softmax output of a small and shallow student model match the large and deep teacher model. In our experiments, we evaluate knowledge distillation training method for remote sensing scene classification on four public datasets: AID dataset, UCMerced dataset, NWPU-RESISC dataset, and EuroSAT dataset. Results show that our proposed training method was effective and increased overall accuracy (3% in AID experiments, 5% in UCMerced experiments, 1% in NWPU-RESISC and EuroSAT experiments for small and shallow models. We further explored the performance of the student model on small and unbalanced datasets. Our findings indicate that knowledge distillation can improve the performance of small network models on datasets with lower spatial resolution images, numerous categories, as well as fewer training samples.

Optimal artificial neural network architecture selection for performance prediction of compact heat exchanger with the EBaLM-OTR technique

Energy Technology Data Exchange (ETDEWEB)

Wijayasekara, Dumidu, E-mail: wija2589@vandals.uidaho.edu [Department of Computer Science, University of Idaho, 1776 Science Center Drive, Idaho Falls, ID 83402 (United States); Manic, Milos [Department of Computer Science, University of Idaho, 1776 Science Center Drive, Idaho Falls, ID 83402 (United States); Sabharwall, Piyush [Idaho National Laboratory, Idaho Falls, ID (United States); Utgikar, Vivek [Department of Chemical Engineering, University of Idaho, Idaho Falls, ID 83402 (United States)

2011-07-15

Highlights: > Performance prediction of PCHE using artificial neural networks. > Evaluating artificial neural network performance for PCHE modeling. > Selection of over-training resilient artificial neural networks. > Artificial neural network architecture selection for modeling problems with small data sets. - Abstract: Artificial Neural Networks (ANN) have been used in the past to predict the performance of printed circuit heat exchangers (PCHE) with satisfactory accuracy. Typically published literature has focused on optimizing ANN using a training dataset to train the network and a testing dataset to evaluate it. Although this may produce outputs that agree with experimental results, there is a risk of over-training or over-learning the network rather than generalizing it, which should be the ultimate goal. An over-trained network is able to produce good results with the training dataset but fails when new datasets with subtle changes are introduced. In this paper we present EBaLM-OTR (error back propagation and Levenberg-Marquardt algorithms for over training resilience) technique, which is based on a previously discussed method of selecting neural network architecture that uses a separate validation set to evaluate different network architectures based on mean square error (MSE), and standard deviation of MSE. The method uses k-fold cross validation. Therefore in order to select the optimal architecture for the problem, the dataset is divided into three parts which are used to train, validate and test each network architecture. Then each architecture is evaluated according to their generalization capability and capability to conform to original data. The method proved to be a comprehensive tool in identifying the weaknesses and advantages of different network architectures. The method also highlighted the fact that the architecture with the lowest training error is not always the most generalized and therefore not the optimal. Using the method the testing

Optimal artificial neural network architecture selection for performance prediction of compact heat exchanger with the EBaLM-OTR technique

International Nuclear Information System (INIS)

Wijayasekara, Dumidu; Manic, Milos; Sabharwall, Piyush; Utgikar, Vivek

2011-01-01

Highlights: → Performance prediction of PCHE using artificial neural networks. → Evaluating artificial neural network performance for PCHE modeling. → Selection of over-training resilient artificial neural networks. → Artificial neural network architecture selection for modeling problems with small data sets. - Abstract: Artificial Neural Networks (ANN) have been used in the past to predict the performance of printed circuit heat exchangers (PCHE) with satisfactory accuracy. Typically published literature has focused on optimizing ANN using a training dataset to train the network and a testing dataset to evaluate it. Although this may produce outputs that agree with experimental results, there is a risk of over-training or over-learning the network rather than generalizing it, which should be the ultimate goal. An over-trained network is able to produce good results with the training dataset but fails when new datasets with subtle changes are introduced. In this paper we present EBaLM-OTR (error back propagation and Levenberg-Marquardt algorithms for over training resilience) technique, which is based on a previously discussed method of selecting neural network architecture that uses a separate validation set to evaluate different network architectures based on mean square error (MSE), and standard deviation of MSE. The method uses k-fold cross validation. Therefore in order to select the optimal architecture for the problem, the dataset is divided into three parts which are used to train, validate and test each network architecture. Then each architecture is evaluated according to their generalization capability and capability to conform to original data. The method proved to be a comprehensive tool in identifying the weaknesses and advantages of different network architectures. The method also highlighted the fact that the architecture with the lowest training error is not always the most generalized and therefore not the optimal. Using the method the

Enlarge the training set based on inter-class relationship for face recognition from one image per person.

Directory of Open Access Journals (Sweden)

Qin Li

Full Text Available In some large-scale face recognition task, such as driver license identification and law enforcement, the training set only contains one image per person. This situation is referred to as one sample problem. Because many face recognition techniques implicitly assume that several (at least two images per person are available for training, they cannot deal with the one sample problem. This paper investigates principal component analysis (PCA, Fisher linear discriminant analysis (LDA, and locality preserving projections (LPP and shows why they cannot perform well in one sample problem. After that, this paper presents four reasons that make one sample problem itself difficult: the small sample size problem; the lack of representative samples; the underestimated intra-class variation; and the overestimated inter-class variation. Based on the analysis, this paper proposes to enlarge the training set based on the inter-class relationship. This paper also extends LDA and LPP to extract features from the enlarged training set. The experimental results show the effectiveness of the proposed method.

The connection-set algebra--a novel formalism for the representation of connectivity structure in neuronal network models.

Science.gov (United States)

Djurfeldt, Mikael

2012-07-01

The connection-set algebra (CSA) is a novel and general formalism for the description of connectivity in neuronal network models, from small-scale to large-scale structure. The algebra provides operators to form more complex sets of connections from simpler ones and also provides parameterization of such sets. CSA is expressive enough to describe a wide range of connection patterns, including multiple types of random and/or geometrically dependent connectivity, and can serve as a concise notation for network structure in scientific writing. CSA implementations allow for scalable and efficient representation of connectivity in parallel neuronal network simulators and could even allow for avoiding explicit representation of connections in computer memory. The expressiveness of CSA makes prototyping of network structure easy. A C+ + version of the algebra has been implemented and used in a large-scale neuronal network simulation (Djurfeldt et al., IBM J Res Dev 52(1/2):31-42, 2008b) and an implementation in Python has been publicly released.

Acute Effects of Back Squats on Countermovement Jump Performance Across Multiple Sets of A Contrast Training Protocol in Resistance-Trained Males.

Science.gov (United States)

Bauer, Pascal; Sansone, Pierpaolo; Mitter, Benedikt; Makivic, Bojan; Seitz, Laurent B; Tschan, Harald

2018-01-03

The present study was designed to evaluate the voluntary post-activation potentiation (PAP) effects of moderate (MI) or high intensity (HI) back squat exercises on countermovement jump (CMJ) performance across multiple sets of a contrast training protocol. Sixty resistance-trained male subjects (age, 23.3 ± 3.3 y; body mass, 86.0 ± 13.9 kg; parallel back squat 1-repetition maximum [1-RM], 155.2 ± 30.0 kg) participated in a randomized, cross-over study. After familiarization, the subjects visited the laboratory on three separate occasions. They performed a contrast PAP protocol comprising three sets of either MI (6×60% of 1-RM) or HI back squats (4x90% of 1-RM) or 20 s of recovery (CTRL) alternated with seven CMJs that were performed at 15 s, and 1, 3, 5, 7, 9 and 11 min after the back squats or recovery. Jump height and relative peak power output recorded with a force platform during MI and HI conditions were compared to those recorded during control condition to calculate the voluntary PAP effect. CMJ performance was decreased immediately after the squats but increased across all three sets of MI and HI between 3 - 7 minutes post-recovery. However, voluntary PAP effects were small or trivial and no difference between the three sets could be found. These findings demonstrate that practitioners can use MI and HI back squats to potentiate CMJs across a contrast training protocol, but a minimum of 3 min of recovery after the squats is needed to benefit from voluntary PAP.

Methodical approach to training of IT-professionals based on networking

Directory of Open Access Journals (Sweden)

Vyacheslav V. Zolotarev

2017-12-01

Full Text Available Increasing requirements to the content and form of higher education in conditions of digital economy set new tasks for professors: formation of applied competences, the involvement of students in project activities, provision of students’ online support, their individual and project work. The growing load on university professors complicates satisfaction of these requirements. The development of the professors’ network interaction makes it possible to redistribute the load for disciplines methodological provision. The article reveals possibilities of professors’ network interaction by using innovative teaching methods including gaming forms and online courses. The research scientific novelty is to implement the professors’ network interaction and experimental application of innovative teaching methods. Network interaction was carried out through the educational process of students’ preparation in following areas: information security, applied information technology, business informatics.

Applying neural networks to control the TFTR neutral beam ion sources

International Nuclear Information System (INIS)

Lagin, L.

1992-01-01

This paper describes the application of neural networks to the control of the neutral beam long-pulse positive ion source accelerators on the Tokamak Fusion Test Reactor (TFTR) at Princeton University. Neural networks were used to learn how the operators adjust the control setpoints when running these sources. The data sets used to train these networks were derived from a large database containing actual setpoints and power supply waveform calculations for the 1990 run period. The networks learned what the optimum control setpoints should initially be set based uon desired accel voltage and perveance levels. Neural networks were also used to predict the divergence of the ion beam

Vision-based mobile robot navigation through deep convolutional neural networks and end-to-end learning

Science.gov (United States)

Zhang, Yachu; Zhao, Yuejin; Liu, Ming; Dong, Liquan; Kong, Lingqin; Liu, Lingling

2017-09-01

In contrast to humans, who use only visual information for navigation, many mobile robots use laser scanners and ultrasonic sensors along with vision cameras to navigate. This work proposes a vision-based robot control algorithm based on deep convolutional neural networks. We create a large 15-layer convolutional neural network learning system and achieve the advanced recognition performance. Our system is trained from end to end to map raw input images to direction in supervised mode. The images of data sets are collected in a wide variety of weather conditions and lighting conditions. Besides, the data sets are augmented by adding Gaussian noise and Salt-and-pepper noise to avoid overfitting. The algorithm is verified by two experiments, which are line tracking and obstacle avoidance. The line tracking experiment is proceeded in order to track the desired path which is composed of straight and curved lines. The goal of obstacle avoidance experiment is to avoid the obstacles indoor. Finally, we get 3.29% error rate on the training set and 5.1% error rate on the test set in the line tracking experiment, 1.8% error rate on the training set and less than 5% error rate on the test set in the obstacle avoidance experiment. During the actual test, the robot can follow the runway centerline outdoor and avoid the obstacle in the room accurately. The result confirms the effectiveness of the algorithm and our improvement in the network structure and train parameters

Illusory Motion Reproduced by Deep Neural Networks Trained for Prediction

Directory of Open Access Journals (Sweden)

Eiji Watanabe

2018-03-01

Full Text Available The cerebral cortex predicts visual motion to adapt human behavior to surrounding objects moving in real time. Although the underlying mechanisms are still unknown, predictive coding is one of the leading theories. Predictive coding assumes that the brain's internal models (which are acquired through learning predict the visual world at all times and that errors between the prediction and the actual sensory input further refine the internal models. In the past year, deep neural networks based on predictive coding were reported for a video prediction machine called PredNet. If the theory substantially reproduces the visual information processing of the cerebral cortex, then PredNet can be expected to represent the human visual perception of motion. In this study, PredNet was trained with natural scene videos of the self-motion of the viewer, and the motion prediction ability of the obtained computer model was verified using unlearned videos. We found that the computer model accurately predicted the magnitude and direction of motion of a rotating propeller in unlearned videos. Surprisingly, it also represented the rotational motion for illusion images that were not moving physically, much like human visual perception. While the trained network accurately reproduced the direction of illusory rotation, it did not detect motion components in negative control pictures wherein people do not perceive illusory motion. This research supports the exciting idea that the mechanism assumed by the predictive coding theory is one of basis of motion illusion generation. Using sensory illusions as indicators of human perception, deep neural networks are expected to contribute significantly to the development of brain research.

LAI inversion from optical reflectance using a neural network trained with a multiple scattering model

Science.gov (United States)

Smith, James A.

1992-01-01

The inversion of the leaf area index (LAI) canopy parameter from optical spectral reflectance measurements is obtained using a backpropagation artificial neural network trained using input-output pairs generated by a multiple scattering reflectance model. The problem of LAI estimation over sparse canopies (LAI 1000 percent for low LAI. Minimization methods applied to merit functions constructed from differences between measured reflectances and predicted reflectances using multiple-scattering models are unacceptably sensitive to a good initial guess for the desired parameter. In contrast, the neural network reported generally yields absolute percentage errors of <30 percent when weighting coefficients trained on one soil type were applied to predicted canopy reflectance at a different soil background.

Predictive Control of Hydronic Floor Heating Systems using Neural Networks and Genetic Algorithms

DEFF Research Database (Denmark)

Vinther, Kasper; Green, Torben; Østergaard, Søren

2017-01-01

This paper presents the use a neural network and a micro genetic algorithm to optimize future set-points in existing hydronic floor heating systems for improved energy efficiency. The neural network can be trained to predict the impact of changes in set-points on future room temperatures. Additio...... space is not guaranteed. Evaluation of the performance of multiple neural networks is performed, using different levels of information, and optimization results are presented on a detailed house simulation model....

Path selection rules for droplet trains in single-lane microfluidic networks

Science.gov (United States)

Amon, A.; Schmit, A.; Salkin, L.; Courbin, L.; Panizza, P.

2013-07-01

We investigate the transport of periodic trains of droplets through microfluidic networks having one inlet, one outlet, and nodes consisting of T junctions. Variations of the dilution of the trains, i.e., the distance between drops, reveal the existence of various hydrodynamic regimes characterized by the number of preferential paths taken by the drops. As the dilution increases, this number continuously decreases until only one path remains explored. Building on a continuous approach used to treat droplet traffic through a single asymmetric loop, we determine selection rules for the paths taken by the drops and we predict the variations of the fraction of droplets taking these paths with the parameters at play including the dilution. Our results show that as dilution decreases, the paths are selected according to the ascending order of their hydrodynamic resistance in the absence of droplets. The dynamics of these systems controlled by time-delayed feedback is complex: We observe a succession of periodic regimes separated by a wealth of bifurcations as the dilution is varied. In contrast to droplet traffic in single asymmetric loops, the dynamical behavior in networks of loops is sensitive to initial conditions because of extra degrees of freedom.

Hematopoietic cell transplantation: Training challenges and potential opportunities through networking and integration of modern technologies to the practice setting.

Science.gov (United States)

Kharfan-Dabaja, Mohamed A; Aljurf, Mahmoud

2017-12-01

Hematopoietic cell transplantation (HCT), particularly allogeneic HCT, is a complex and a high-risk procedure requiring expertise to manage potential treatment complications. Published data supports the value of quality management systems in improving post-transplant outcomes; however, there are no universally established, or agreed upon, criteria to assess adequacy of training of physicians, transplant or nontransplant, and supporting staff, among others. It is of paramount importance for transplant centers to identify the needed area(s) of expertise in order to seek appropriate training for their staff. Moreover, transplant physicians need to keep up-to-date with the rapidly occurring advances in the field. Outcomes of patients undergoing HCT are affected by various factors related to patient, disease, procedure, preventative, and supportive strategies, among others. Accordingly, availability of databases is necessary to collect information on these variables and use to benchmark future prospective clinical trials aiming at further improving clinical outcomes. Twinning with leading centers worldwide is helping to not only bridge the survival gap of patients diagnosed with cancer in the developing vis-à-vis the developed world, but eventually closing it. The advent of the World Wide Web and revolution in telecommunication has made access to information more readily available to various sectors including healthcare. Telemedicine is enabling healthcare delivery to remote and underserved geographic areas. In the setting of HCT, ensuring compliance to prescribed therapies and post-transplant surveillance are some areas where implementing telemedicine programs could fulfill an unmet need. Copyright © 2017 King Faisal Specialist Hospital & Research Centre. Published by Elsevier B.V. All rights reserved.

Asian network for education in nuclear technology: An initiative to promote education and training in nuclear technology

International Nuclear Information System (INIS)

Kosilov, A.

2006-01-01

It has become increasingly clear that there is a need to consolidate the efforts of academia and industry in education and training. Partnerships of operating organizations with educational institutions and universities that provide qualified professionals for the nuclear industry should be assessed based upon medium and long term needs and strengthened where needed. In this regard the IAEA is taking the necessary action to initiate this kind of partnership through continuous networking. The paper describes the IAEA approach to promoting education and training through the Asian Network for Education in Nuclear Technology (ANENT). (author)

A proposal to order the neutron data set in neutron spectrometry using the RDANN methodology

Energy Technology Data Exchange (ETDEWEB)

Ortiz R, J.M.; Martinez B, M.R.; Vega C, H.R. [UAZ, Av. Ramon Lopez Velarde No. 801, 98000 Zacatecas (Mexico)

2006-07-01

A new proposal to order a neutron data set in the design process of artificial neural networks in the neutron spectrometry field is presented for first time. The robust design of artificial neural networks methodology was applied to 187 neutron spectra data set compiled by the International Atomic Energy Agency. Four cases of grouping the neutron spectra were considered and around 1000 different neural networks were designed, trained and tested with different net topologies each one. After carrying out the systematic methodology for all the cases, it was determined that the best neural network topology that produced the best reconstructed neutron spectra was case with 187 neutron spectra data set, determining that the best neural network topology is: 7 entrance neurons, 14 neurons in a hidden layer and 31 neurons in the exit layer, with a value of 0.1 in the learning rate and 0.1 in the moment. (Author)

A proposal to order the neutron data set in neutron spectrometry using the RDANN methodology

International Nuclear Information System (INIS)

Ortiz R, J.M.; Martinez B, M.R.; Vega C, H.R.

2006-01-01

A new proposal to order a neutron data set in the design process of artificial neural networks in the neutron spectrometry field is presented for first time. The robust design of artificial neural networks methodology was applied to 187 neutron spectra data set compiled by the International Atomic Energy Agency. Four cases of grouping the neutron spectra were considered and around 1000 different neural networks were designed, trained and tested with different net topologies each one. After carrying out the systematic methodology for all the cases, it was determined that the best neural network topology that produced the best reconstructed neutron spectra was case with 187 neutron spectra data set, determining that the best neural network topology is: 7 entrance neurons, 14 neurons in a hidden layer and 31 neurons in the exit layer, with a value of 0.1 in the learning rate and 0.1 in the moment. (Author)

Predicting Positive and Negative Relationships in Large Social Networks.

Directory of Open Access Journals (Sweden)

Guan-Nan Wang

Full Text Available In a social network, users hold and express positive and negative attitudes (e.g. support/opposition towards other users. Those attitudes exhibit some kind of binary relationships among the users, which play an important role in social network analysis. However, some of those binary relationships are likely to be latent as the scale of social network increases. The essence of predicting latent binary relationships have recently began to draw researchers' attention. In this paper, we propose a machine learning algorithm for predicting positive and negative relationships in social networks inspired by structural balance theory and social status theory. More specifically, we show that when two users in the network have fewer common neighbors, the prediction accuracy of the relationship between them deteriorates. Accordingly, in the training phase, we propose a segment-based training framework to divide the training data into two subsets according to the number of common neighbors between users, and build a prediction model for each subset based on support vector machine (SVM. Moreover, to deal with large-scale social network data, we employ a sampling strategy that selects small amount of training data while maintaining high accuracy of prediction. We compare our algorithm with traditional algorithms and adaptive boosting of them. Experimental results of typical data sets show that our algorithm can deal with large social networks and consistently outperforms other methods.

Predicting Positive and Negative Relationships in Large Social Networks.

Science.gov (United States)

Wang, Guan-Nan; Gao, Hui; Chen, Lian; Mensah, Dennis N A; Fu, Yan

2015-01-01

In a social network, users hold and express positive and negative attitudes (e.g. support/opposition) towards other users. Those attitudes exhibit some kind of binary relationships among the users, which play an important role in social network analysis. However, some of those binary relationships are likely to be latent as the scale of social network increases. The essence of predicting latent binary relationships have recently began to draw researchers' attention. In this paper, we propose a machine learning algorithm for predicting positive and negative relationships in social networks inspired by structural balance theory and social status theory. More specifically, we show that when two users in the network have fewer common neighbors, the prediction accuracy of the relationship between them deteriorates. Accordingly, in the training phase, we propose a segment-based training framework to divide the training data into two subsets according to the number of common neighbors between users, and build a prediction model for each subset based on support vector machine (SVM). Moreover, to deal with large-scale social network data, we employ a sampling strategy that selects small amount of training data while maintaining high accuracy of prediction. We compare our algorithm with traditional algorithms and adaptive boosting of them. Experimental results of typical data sets show that our algorithm can deal with large social networks and consistently outperforms other methods.

Control of autonomous robot using neural networks

Science.gov (United States)

Barton, Adam; Volna, Eva

2017-07-01

The aim of the article is to design a method of control of an autonomous robot using artificial neural networks. The introductory part describes control issues from the perspective of autonomous robot navigation and the current mobile robots controlled by neural networks. The core of the article is the design of the controlling neural network, and generation and filtration of the training set using ART1 (Adaptive Resonance Theory). The outcome of the practical part is an assembled Lego Mindstorms EV3 robot solving the problem of avoiding obstacles in space. To verify models of an autonomous robot behavior, a set of experiments was created as well as evaluation criteria. The speed of each motor was adjusted by the controlling neural network with respect to the situation in which the robot was found.