A First in Combining Science Discovery Technologies: Federated Search and Speech-Indexed Multimedia

A First in Combining Science Discovery Technologies:
Federated Search and Speech-Indexed Multimedia

Oak Ridge, TN - The DOE Office of Scientific and Technical Information (OSTI) announced today a new tool in scientific discovery technology. Now citizens and researchers alike can search for both written and spoken words in a whole range of media using OSTI's new, speech-indexed multimedia within large scientific search portals. To this point, online searches for scientific information have been limited to text, such as within scientific papers. The new development uses unique speech-recognition search technology in combination with OSTI's two federated search portals, ScienceAcceleator.gov and WorldWideScience.org, which search a wide range of DOE and worldwide databases, respectively. This vastly extends the reach of federated searching and could lead to new connections and new breakthroughs.

OSTI Director Dr. Walter Warnick said, “The addition of speech-indexed multimedia to federated search is a major milestone in advancing scientific discovery, especially as R&D results are increasingly recorded in video, audio, animation, and visualization.”

OSTI has pioneered the use of federated searching to enable the science community to search and access large, decentralized collections of scientific and technical information.  Major federated search products include ScienceAccelerator.gov, Science.gov, and WorldWideScience.org.  ScienceAccelerator.gov was developed by OSTI to search databases covering DOE research and research of interest to DOE. Science.gov, operated by OSTI on behalf of the Science.gov Alliance, searches the science databases and websites of fourteen U.S. federal agencies, while WorldWideScience.org - operated by OSTI on behalf of the WorldWideScience Alliance - searches scientific collections of the U.S. and over 70 other nations. 

OSTI partnered with the Microsoft Research Audio Video Indexing System (MAVIS) project to build a multimedia search engine called ScienceCinema.  Introduced in February 2011 with approximately 1,000 scientific videos from U.S. DOE national laboratories, ScienceCinema content continues to grow with the recent initial installment from the multimedia collection of CERN, the European Organization for Nuclear Research. 

Along with huge DOE R&D collections such as the Information Bridge, DOepatents, Energy Citations Database, and DOE R&D Accomplishments, ScienceAccelerator.gov will now search the multimedia of ScienceCinema. 

The addition of speech-indexed multimedia searching through WorldWideScience.org was announced by Dr. Warnick in his presentation to the International Council for Scientific and Technical Information (ICSTI), which serves as primary sponsor for WorldWideScience.org.  Dr. Warnick also announced two other major enhancements to WorldWideScience.org:  the addition of Arabic as the tenth language to be part of its multilingual translations capability and the introduction of a mobile web version, http://m.worldwidescience.org (another first in federated search technology).

OSTI, within the DOE Office of Science, is responsible for making R&D findings available and useful to researchers and the public.