Committee-Based Active Learning for Speech Recognition

Yuzo HAMANAKA  Koichi SHINODA  Takuya TSUTAOKA  Sadaoki FURUI  Tadashi EMORI  Takafumi KOSHINAKA  

IEICE TRANSACTIONS on Information and Systems   Vol.E94-D   No.10   pp.2015-2023
Publication Date: 2011/10/01
Online ISSN: 1745-1361
DOI: 10.1587/transinf.E94.D.2015
Print ISSN: 0916-8532
Type of Manuscript: PAPER
Category: Speech and Hearing
active learning,  query by committee,  LVCSR,  progressive alignment,  

Full Text: PDF(584KB)>>
Buy this Article

We propose a committee-based method of active learning for large vocabulary continuous speech recognition. Multiple recognizers are trained in this approach, and the recognition results obtained from these are used for selecting utterances. Those utterances whose recognition results differ the most among recognizers are selected and transcribed. Progressive alignment and voting entropy are used to measure the degree of disagreement among recognizers on the recognition result. Our method was evaluated by using 191-hour speech data in the Corpus of Spontaneous Japanese. It proved to be significantly better than random selection. It only required 63 h of data to achieve a word accuracy of 74%, while standard training (i.e., random selection) required 103 h of data. It also proved to be significantly better than conventional uncertainty sampling using word posterior probabilities.