Title :
The use of a linguistically motivated language model in conversational speech recognition
Author :
Wang, Wen ; Stolcke, Andreas ; Harper, Mary P.
Author_Institution :
Speech Technol. & Res. Lab., SRI Int., Menlo Park, CA, USA
Abstract :
Structured language models have recently been shown to give significant improvements in large-vocabulary recognition relative to traditional word N-gram models, but typically imply a heavy computational burden and have not been applied to large training sets or complex recognition systems. Previously, we developed a linguistically motivated and computationally efficient almost-parsing language model, using a data structure derived from constraint dependency grammar parsing, that tightly integrates knowledge of words, lexical features, and syntactic constraints. We show that such a model can be used effectively and efficiently in all stages of a complex, multi-pass conversational telephone speech recognition system. Compared to a state-of-the-art 4-gram interpolated word- and class-based language model, we obtained a 6.2% relative word error reduction (a 1.6% absolute reduction) on a recent NIST evaluation set.
Keywords :
learning (artificial intelligence); natural languages; speech recognition; constraint dependency grammar parsing; conversational speech recognition; data structure; lexical features; linguistically motivated language model; structured language models; syntactic constraints; telephone speech recognition system; word N-gram models; word error reduction; Data structures; Decoding; Laboratories; Lattices; Mel frequency cepstral coefficient; NIST; Natural languages; Speech recognition; Telephony; Training data;
Conference_Titel :
Acoustics, Speech, and Signal Processing, 2004. Proceedings. (ICASSP '04). IEEE International Conference on
Print_ISBN :
0-7803-8484-9
DOI :
10.1109/ICASSP.2004.1325972