Nonnative speech recognition based on bilingual model modification at state level

Qingqing Zhang, Jielin Pan, Shui Duen Chan, Yonghong Yan

Research output: Chapter in book / Conference proceedingConference article published in proceeding or bookAcademic researchpeer-review

Abstract

This paper presents a novel bilingual model modification approach to improve nonnative speech recognition accuracy when the variations of accented pronunciations occur. Each state of baseline nonnative acoustic model is modified with several candidate states from the auxiliary acoustic model, which is trained on speakers’ mother language. State mapping criterion and n-best candidates are investigated, and different numbers of Gaussian mixtures of the auxiliary acoustic model are compared based on a grammarconstrained speech recognition system. Using this bilingual model modification approach, compared to the nonnative acoustic model which has already been well trained by adaptation technique MAP, the Phrase Error Rate further achieves a 5.83% relative reduction, while only a small relative increase on Real Time Factor occurs.
Original languageEnglish
Title of host publicationAdvances in Intelligent and Soft Computing
PublisherSpringer Verlag
Pages299-309
Number of pages11
ISBN (Electronic)9783642012150
Publication statusPublished - 1 Jan 2009
Event6th International Symposium of Neural Networks, ISNN 2009 - Wuhan, China
Duration: 26 May 200929 May 2009

Publication series

NameAdvances in Intelligent and Soft Computing
Volume56
ISSN (Print)1867-5662
ISSN (Electronic)1860-0794

Conference

Conference6th International Symposium of Neural Networks, ISNN 2009
CountryChina
CityWuhan
Period26/05/0929/05/09

ASJC Scopus subject areas

  • Computer Science(all)

Cite this