Homophone words is one of the specific problems of Automatic Speech Recognition (ASR) in French. Moreover, this phenomenon is particularly high for some inflections like the singular/plural inflection (72% of the 40.7K lemma of our 240K word dictionary have inflected forms which are homophonic). In order to take into account word-dependencies spanning over a variable number of words, it is interesting to merge local language models, like 3-gram or 3-class models, with large-span models. We present in this paper two kinds of models : a phrase-based model, using phrases obtained from a training corpus by means of a finitestate parser; a homophone cache-based model, using derivation of constraints from word histories stored in a cache memory.
Cite as: Béchet, F., Nasr, A., Spriet, T., Mori, R.d. (1999) Large Span statistical language models: application to homophone disambiguation for large vocabulary speech recognition in French. Proc. 6th European Conference on Speech Communication and Technology (Eurospeech 1999), 1763-1766, doi: 10.21437/Eurospeech.1999-352
@inproceedings{bechet99_eurospeech,
author={Frédéric Béchet and Alexis Nasr and Thierry Spriet and Renato de Mori},
title={{Large Span statistical language models: application to homophone disambiguation for large vocabulary speech recognition in French}},
year=1999,
booktitle={Proc. 6th European Conference on Speech Communication and Technology (Eurospeech 1999)},
pages={1763--1766},
doi={10.21437/Eurospeech.1999-352}
}