Incremental language models for speech recognition using finite-state transducers
Tóm tắt
In the context of the weighted finite-state transducer approach to speech recognition, we investigate a novel decoding strategy to deal with very large n-gram language models often used in large-vocabulary systems. In particular, we present an alternative to full, static expansion and optimization of the finite-state transducer network. This alternative is useful when the individual knowledge sources, modeled as transducers, are too large to be composed and optimized. While the recognition decoder perceives a single, weighted finite-state transducer, we apply a divide-and-conquer technique to split the language model into two parts which add up exactly to the original language model. We investigate the merits of these 'incremental language models' and present some initial results.
Từ khóa
#Natural languages #Speech recognition #Decoding #Hidden Markov models #Acoustic transducers #Laboratories #Context modeling #Oceans #Surveillance #OxygenTài liệu tham khảo
aubert, 1999, One pass cross word decoding for large vocabularies based on a fexical tree search organization, Proc European Conference on Speech Communication and Technology Budapest, 4, 1559
mohri, 2000, Weighted finite-state transducers in speech recognition, Proc Automatic Speech Recognition workshop 2000, 1, 97
klakow, 1998, Language model investigations related to broadcast news, Proc DARPA Workshop on Automatic Transcription of Broadcast News
ortmanns, 1996, Languae model lookahead for large vocabulary speech recognition, Proc International Conference on Spoken Language Processing, 2095
10.1109/ICASSP.1994.389702
10.1109/89.817460
10.1109/ICASSP.1999.758062
liolje, 2000, The AT&T LVCSR-2000 System, Proc of The NIST Large Vocabulary Conversational Speech Recognition Workshop
hetherington, 2001, An efficient implementation of phonological rules using finite-state transducers, Proc European Conference on Speech Communication and Technology
aubert, 2000, A brief overview of decoding techniques for large vocabulary continuous speech recognition, Proc Automatic Speech Recognition workshop 2000, 1, 91