| Paper: | 03-P1.8 (ICASSP 2003 Paper) | ||
| Session: | ICASSP 2003 Papers | ||
| Time: | Tuesday, May 18, 13:00 - 15:00 | ||
| Presentation: | Poster (ICASSP 2003 Presentation) | ||
| Topic: | ICASSP 2003 Papers: ICASSP 2003 Papers | ||
| Title: | PERVASIVE UNSUPERVISED ADAPTATION FOR LECTURE SPEECH TRANSCRIPTION | ||
| Authors: | Daniel Willett; NTT Corporation | ||
| Thomas Niesler; NTT Corporation | |||
| Erik McDermott; NTT Corporation | |||
| Yasuhiro Minami; NTT Corporation | |||
| Shigeru Katagiri; NTT Corporation | |||
| Abstract: | Unsupervised adaptation has evolved as a popular approach for tuning the acoustic models of speaker-independent speech recognition systems tospecific speakers, speaker groups or channel conditions while making use of untranscribed data only.This work focuses on procedures for unsupervised adaptation of other probabilistic models that are involved in state-of-the-art speechrecognizers and on the joint adaptation of multiple knowledge sources. In particular, we outline and evaluate approaches for adaptingthe language model and the pronunciation model (lexicon) without supervision. Initial experiments on off-line lecture speech transcription achieved littlebut promising word error rate improvements with each approach applied separately. The experimental results on the joint application of acoustic,language and pronunciation model adaptation indicate that the individually achievable performance improvements are additive. | ||
| Back | |||
Home -||-
Organizing Committee -||-
Technical Committee -||-
Technical Program -||-
Plenaries
Paper Submission -||-
Special Sessions -||-
ITT -||-
Paper Review -||-
Exhibits -||-
Tutorials
Information -||-
Registration -||-
Travel Insurance -||-
Housing -||-
Workshops