Merging of native and non-native speech for low-resource accented ASR

This paper presents our recent study on low-resource automatic speech recognition (ASR) system with accented speech. We propose multi-accent Subspace Gaussian Mixture Models (SGMM) and accent-specific Deep Neural Networks (DNN) for improving non-native ASR performance. In the SGMM framework, we pres...

詳細記述

保存先:
書誌詳細
主要な著者: Samson Juan, Sarah, Besacier, Laurent, Lecouteux, Benjamin, Tien-Ping, Tan
フォーマット: E-Article
言語:English
出版事項: Springer Verlag 2015
主題:
オンライン・アクセス:http://ir.unimas.my/id/eprint/12098/1/No%2035%20%28abstrak%29.pdf
http://ir.unimas.my/id/eprint/12098/
http://www.scopus.com/inward/record.url?eid=2-s2.0-84952362047&partnerID=40&md5=6bc512988afc29cd7ca4af16a836f0b3
タグ: タグ追加
タグなし, このレコードへの初めてのタグを付けませんか!
その他の書誌記述
要約:This paper presents our recent study on low-resource automatic speech recognition (ASR) system with accented speech. We propose multi-accent Subspace Gaussian Mixture Models (SGMM) and accent-specific Deep Neural Networks (DNN) for improving non-native ASR performance. In the SGMM framework, we present an original language weighting strategy to merge the globally shared parameters of two models based on native and non-native speech espectively. In the DNN framework, a native deep neural net is fine-tuned to non-native speech. Over the non-native baseline, we achieved relative improvement of 15% for multi-accent SGMM and 34% for accent-specific DNN with speaker adaptation.