Accessing Information in Spoken Audio

April 19-20, 1999
Cambridge, UK

The THISL Broadcast News Retrieval System

Dave Abberley (1), David Kirby (2), Steve Renals (1), and Tony Robinson (3)

(1) University of Sheffield, Department of Computer Science, UK
(2) BBC, Research and Development, UK
(3) SoftSound, UK

This paper described the THISL spoken document retrieval system for British and North American Broadcast News. The system is based on the Abbot large vocabulary speech recognizer, using a recurrent network acoustic model, and a probabilistic text retrieval system. We discuss the development of a realtime British English Broadcast News system, and its integration into a spoken document retrieval system. Detailed evaluation is performed using a similar North American Broadcast News system, to take advantage of the TREC SDR evaluation methodology. We report results on this evaluation, with particular reference to the effect of query expansion and of automatic segmentation algorithms.


Full Paper (PDF)   Full Paper (Zipped Postscript)

Bibliographic reference.  Abberley, Dave / Kirby, David / Renals, Steve / Robinson, Tony (1999): "The THISL broadcast news retrieval system ", In Access-Audio-1999, 14-19.