OnTAP: mixed-media multi-lingual language processing
Sean Colbath, Scott Shepard, Daben Liu, Amit Kumar Srivastava, Jayadev Billa, James Gugerty, Daniel Kiecza, Mohammed Noamany, Kathleen M. Egan, Francis Kubala · International Conference on Human Language Technology Research · 2002
OnTAP (Online Text and Audio Processing) is a fully automatic language processing demonstration system for processing human-language data in mixed media and multiple languages. Created under the DARPA TIDES program, OnTAP acts as a technology test bed for the program. The core technologies that comprise OnTAP include: automatic speech recognition in multiple languages, speaker change detection, speaker clustering, speaker identification, named entity extraction, topic identification, story segmentation, fact extraction, event detection, and event tracking. OnTAP includes the Oasis Audio Indexing system for processing audio from radio and TV.