DMCA
The Thisl Spoken Document Retrieval System (1998)
BibTeX
@MISC{Steve98thethisl,
author = {Dave Abberley Steve and Steve Renals and Gary Cook and Tony Robinson},
title = {The Thisl Spoken Document Retrieval System},
year = {1998}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
INTRODUCTION The THISL spoken document retrieval system is based on the ABBOT Large Vocabulary Continuous Speech Recognition (LVCSR) system developed by Cambridge University, Sheffield University and SoftSound, and uses PRISE (NIST) for indexing and retrieval. We participated in full SDR mode. Our approach was to transcribe the spoken documents at the word level using ABBOT, indexing the resulting text transcriptions using PRISE. The LVCSR system uses a recurrent network-based acoustic model (with no adaptation to different conditions) trained on the 50 hour Broadcast News training set, a 65,000 word vocabulary and a trigram language model derived from Broadcast News text. Words in queries which were out-of-vocabulary (OOV) were word spotted at query time (utilizing the posterior phone probabilities output by the acoustic model), added to the transcriptions of the relevant documents and the collection was then re-indexed. We generated pronunciati
Keyphrases
thisl spoken document retrieval system lvcsr system cambridge university word vocabulary word level posterior phone probability output broadcast news text hour broadcast news training set recurrent network-based acoustic model relevant document sheffield university trigram language model text transcription query time acoustic model full sdr mode spoken document different condition
