Robust language modeling for a small corpus of target tasks using class-combined word statistics and selective use of a general corpus

Systems and Computers in Japan

Volumn 34, Issue 12, 2003, Pages 92-102

  Wada, Yosuke ^a   Kobayashi, Norihiko ^a   Kobayashi, Tetsunori ^a  

^a WASEDA UNIVERSITY   (Japan)

Author keywords

Class N gram; Language model; Large vocabulary continuous speech recognition; Task adaptation

Indexed keywords

COMPUTATIONAL LINGUISTICS; MATHEMATICAL MODELS; STATISTICS; TEXT PROCESSING;

CLASS COMBINED BIGRAM; LANGUAGE MODELS; TASK ADAPTATION;

CONTINUOUS SPEECH RECOGNITION;

EID: 0142091467     PISSN: 08821666     EISSN: None     Source Type: Journal    
DOI: 10.1002/scj.1219     Document Type: Article

References (14)

1. Katz SM. Estimation of probabilities from sparse data for language model component of a speech recognizer. IEEE Trans Acoust Speech Signal Process 1987;ASSP-35:400-401.
2. Witten IH, Bell TC. The zero-frequency problem: Estimating the probabilities of novel events in adaptive text compression. IEEE Trans Inf Theory 1991;37:1085-1094.
3. Niesler TR, Woodland PC. Combination of word-based and category-based language models. Proc ICSLP96, Vol. 1, p 220-223.
4. Samuelsson C, Reichl W. A class-based language model for large-vocabulary speech recognition extracted from part-of-speech statistics. IEEE Proc ICASSP99, Vol. 1, p 537-540.
5. Jelinek F, Mercer FL. Interpolated estimation of Markov source parameters from sparse data. In: Pattern recognition in practice. IEEE; 1980, p 381-397.
6. Hidaki Y, Kobayashi T. Combining word bigram with class bigram to biuld a statistical language model that is robust against topic changes. J Inf Process Soc Japan 1999;98:25-32.
7. Kobayashi N, Kobayashi T. Use of class statistics and word statistics together in constructing a statistical language model for small training data. J Inf Process Soc Japan 1999;99:1-8.
8. Masataki H, Sagisaka Y, Hisaki K, Kawahara T. Task adaptation using MAP estimation in N-gram language modeling. Tech Rep IEICE 1997; SP96-103.
9. Itoh A, Yoshida M. Preliminary task adaptation for conversational speech recognition. Tech Rep IEICE 1996; NLC96-50, SP96-81.
10. Niesler TR, Woodland PC. Combination of word-based and category-based language models. Proc ICSLP96, p 1779-1782, 1997.
11. Ward W, Issar S. A class based language model for speech recognition. IEEE Proc ICASSP96, Vol. 1, p 416-418.
12. Kawahara T, Ri A, Kobayashi T, Takeda K, Minematsu N, Itoh K, Itoh A, Yamamoto M, Yamada A, Utsuro T, Shikano K. Japanese Dictation Software Toolkit (1997 version). J Acoust Soc Japan 1999;55:175-180.
13. Ueberla J. Analysing a simple language model-Some general conclusions for language model for speech recognition. Computer Speech Language 1994;8:153-176.
14. Wada Y, Kobayashi N, Nakano Y, Kobayashi T. Effects of vocabulary expansion through the addition of chained words in large vocabulary continuous speech recognition. J Inf Process Soc Japan 1999;40:1413-1420.