Presentation 2000/1/21
Reading-based Language Model in Japanese Dictation System
Yoshifumi Hirose, Katsunobu Itou, Kiyohiro Shikano, Satoshi Nakamura,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) As Japanese sentences are written using kana and kanji, some words are often written in various ways. In large vocabulary continuous speech recognition, where the vocabulary size is generally restricted, the word writing ambiguity causes the deterioration of the word coverage. This writing ambiguity problem is enlarged by the free creation ability of kanji concatenation in case of person names. In order to improve the word coverage, we propose the following reading-based language models, a reading-based language model, and a kana and kanji mixed language model. The effectiveness for these language models are confirmed by the LVCSR experiments.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) japanese dictation / language model / word coverage / unknown word
Paper # SP99-145
Date of Issue

Conference Information
Committee SP
Conference Date 2000/1/21(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Speech (SP)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Reading-based Language Model in Japanese Dictation System
Sub Title (in English)
Keyword(1) japanese dictation
Keyword(2) language model
Keyword(3) word coverage
Keyword(4) unknown word
1st Author's Name Yoshifumi Hirose
1st Author's Affiliation Graduate School of Information Science, Nara Institute of Science and Technology()
2nd Author's Name Katsunobu Itou
2nd Author's Affiliation Electrotechnical Laboratory
3rd Author's Name Kiyohiro Shikano
3rd Author's Affiliation Graduate School of Information Science, Nara Institute of Science and Technology
4th Author's Name Satoshi Nakamura
4th Author's Affiliation Graduate School of Information Science, Nara Institute of Science and Technology
Date 2000/1/21
Paper # SP99-145
Volume (vol) vol.99
Number (no) 577
Page pp.pp.-
#Pages 9
Date of Issue