%Aigaion2 BibTeX export from Idiap Publications %Thursday 21 November 2024 01:20:46 PM @INPROCEEDINGS{McGreevy04b, author = {McGreevy, Michael}, projects = {Idiap}, month = {12}, title = {Pseudo-Syntactic Language Modeling for Disfluent Speech Recognition}, booktitle = {Proceedings of SST 2004 (10th Australian International Conference on Speech Science & Technology,',','), Sydney, Australia, 2004}, year = {2004}, note = {IDIAP-RR 04-55}, crossref = {mcgreevy04a}, abstract = {Language models for speech recognition are generally trained on text corpora. Since these corpora do not contain the disfluencies found in natural speech, there is a train/test mismatch when these models are applied to conversational speech. In this work we investigate a language model (LM) designed to model these disfluencies as a syntactic process. By modeling self-corrections we obtain an improvement over our baseline syntactic model. We also obtain a 30\\% relative reduction in perplexity from the best performing standard {N-gram} model when we interpolate it with our syntactically derived models.}, pdf = {https://publications.idiap.ch/attachments/papers/2004/mcgreevy-sst04.pdf}, postscript = {ftp://ftp.idiap.ch/pub/papers/speech/mcgreevy-sst04.ps.gz}, ipdmembership={speech}, } crossreferenced publications: @TECHREPORT{McGreevy04a, author = {McGreevy, Michael}, projects = {Idiap}, title = {Pseudo-Syntactic Language Modeling for Disfluent Speech Recognition}, type = {Idiap-RR}, number = {Idiap-RR-55-2004}, year = {2004}, institution = {IDIAP}, note = {Published in Proceedings of SST, 2004}, abstract = {Language models for speech recognition are generally trained on text corpora. Since these corpora do not contain the disfluencies found in natural speech, there is a train/test mismatch when these models are applied to conversational speech. In this work we investigate a language model (LM) designed to model these disfluencies as a syntactic process. By modeling self-corrections we obtain an improvement over our baseline syntactic model. We also obtain a 30\\% relative reduction in perplexity from the best performing standard {N-gram} model when we interpolate it with our syntactically derived models.}, pdf = {https://publications.idiap.ch/attachments/reports/2004/rr04-55.pdf}, postscript = {ftp://ftp.idiap.ch/pub/reports/2004/rr04-55.ps.gz}, ipdmembership={speech}, }