Sixth European Conference on Speech Communication and Technology
In this paper, we describe our recent work in fast automatic transcription of broadcast news programming from radio and television. Given our state-of-the-art BBN BYBLOS primary system  running at 230 times real time (230xRT) we show that eliminating and approximating many computationally expensive components speeds up the system by a factor of more than 20 without significant loss in recognition accuracy. This is accomplished without retraining or changing the baseline system structure. The components of the primary system which are refined include segmentation, adaptation, decoding, crossword rescoring with adaptation, and system combination.
Full Paper (PDF) Gnu-Zipped Postscript
Bibliographic reference. Davenport, Jason / Nguyen, Long / Matsoukas, Spyros / Schwartz, Richard / Makhoul, John (1999): "Toward realtime transcription of broadcast news", In EUROSPEECH'99, 651-654.