9th Annual Conference of the International Speech Communication Association

Brisbane, Australia
September 22-26, 2008

Aggregating Distributed STT, MT, and Information Extraction Engines: The GALE Interoperability-Demo System

John F. Pitrelli (1), Burn L. Lewis (1), Edward A. Epstein (1), Martin Franz (1), Daniel Kiecza (2), Jerome L. Quinn (1), Ganesh Ramaswamy (1), Amit Srivastava (2), Paola Virga (1)

(1) IBM T.J. Watson Research Center, USA; (2) BBN Technologies, USA

Natural-language-processing engines are now attaining accuracy sufficient to begin combining them to perform more complex tasks. The GALE Interoperability Demo system consists of 12 engines including speech recognition, translation, and various information extraction engines, interoperated to make Arabic news video browsable as English text grouped and summarized by topic. Unstructured Information Management Architecture serves as the framework enabling remote pipelined and parallelized operation of these engines operating on their native computing platforms at their home sites.

Full Paper

Bibliographic reference.  Pitrelli, John F. / Lewis, Burn L. / Epstein, Edward A. / Franz, Martin / Kiecza, Daniel / Quinn, Jerome L. / Ramaswamy, Ganesh / Srivastava, Amit / Virga, Paola (2008): "Aggregating distributed STT, MT, and information extraction engines: the GALE interoperability-demo system", In INTERSPEECH-2008, 2743-2746.