このアイテムのアクセス数: 77

このアイテムのファイル:
ファイル 記述 サイズフォーマット 
TASLP.2016.2562505.pdf1.68 MBAdobe PDF見る/開く
完全メタデータレコード
DCフィールド言語
dc.contributor.authorLi, Shengja
dc.contributor.authorAkita, Yuyaja
dc.contributor.authorKawahara, Tatsuyaja
dc.contributor.alternative河原, 達也ja
dc.date.accessioned2017-04-04T04:58:20Z-
dc.date.available2017-04-04T04:58:20Z-
dc.date.issued2016-05-03ja
dc.identifier.issn2329-9290ja
dc.identifier.urihttp://hdl.handle.net/2433/219417-
dc.description.abstractWhile the performance of ASR systems depends on the size of the training data, it is very costly to prepare accurate and faithful transcripts. In this paper, we investigate a semisupervised training scheme, which takes the advantage of huge quantities of unlabeled video lecture archive, particularly for the deep neural network (DNN) acoustic model. In the proposed method, we obtain ASR hypotheses by complementary GMM-and DNN-based ASR systems. Then, a set of CRF-based classifiers is trained to select the correct hypotheses and verify the selected data. The proposed hypothesis combination shows higher quality compared with the conventional system combination method (ROVER). Moreover, compared with the conventional data selection based on confidence measure score, our method is demonstrated more effective for filtering usable data. Significant improvement in the ASR accuracy is achieved over the baseline system and in comparison with the models trained with the conventional system combination and data selection methods.ja
dc.format.mimetypeapplication/pdfja
dc.language.isoengja
dc.publisherInstitute of Electrical and Electronics Engineers Inc.ja
dc.rights©2016 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other users, including reprinting/ republishing this material for advertising or promotional purposes, creating new collective works for resale or redistribution to servers or lists, or reuse of any copyrighted components of this work in other works.ja
dc.rightsThis is not the published version. Please cite only the published version. この論文は出版社版でありません。引用の際には出版社版をご確認ご利用ください。ja
dc.subjectspeech recognitionja
dc.subjectacoustic modelja
dc.subjectsemi-supervised trainingja
dc.subjectlecture transcriptionja
dc.titleSemi-Supervised Acoustic Model Training by Discriminative Data Selection from Multiple ASR Systems' Hypothesesja
dc.type.niitypeJournal Articleja
dc.identifier.jtitleIEEE/ACM Transactions on Audio Speech and Language Processingja
dc.identifier.volume24ja
dc.identifier.issue9-
dc.identifier.spage1524ja
dc.identifier.epage1534ja
dc.relation.doi10.1109/TASLP.2016.2562505ja
dc.textversionauthorja
dc.addressGraduate School of Informatics, Kyoto University-
dc.addressGraduate School of Informatics, Kyoto University-
dc.addressGraduate School of Informatics, Kyoto University-
出現コレクション:学術雑誌掲載論文等

アイテムの簡略レコードを表示する

Export to RefWorks


出力フォーマット 


このリポジトリに保管されているアイテムはすべて著作権により保護されています。