このアイテムのアクセス数: 520

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 
TASL.2006.876727.pdf735.71 kBAdobe 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.authorLane, Ianja
dc.contributor.authorKawahara, Tatsuyaja
dc.contributor.authorMatsui, Tomokoja
dc.contributor.authorNakamura, Satoshija
dc.contributor.alternative河原, 達也ja
dc.date.accessioned2010-10-21T01:45:25Z-
dc.date.available2010-10-21T01:45:25Z-
dc.date.issued2007-01-
dc.identifier.issn1558-7916ja
dc.identifier.urihttp://hdl.handle.net/2433/128902-
dc.description.abstractOne significant problem for spoken language systems is how to cope with users' out-of-domain (OOD) utterances which cannot be handled by the back-end application system. In this paper, we propose a novel OOD detection framework, which makes use of the classification confidence scores of multiple topics and applies a linear discriminant model to perform in-domain verification. The verification model is trained using a combination of deleted interpolation of the in-domain data and minimum-classification-error training, and does not require actual OOD data during the training process, thus realizing high portability. When applied to the "phrasebook" system, a single utterance read-style speech task, the proposed approach achieves an absolute reduction in OOD detection errors of up to 8.1 points (40% relative) compared to a baseline method based on the maximum topic classification score. Furthermore, the proposed approach realizes comparable performance to an equivalent system trained on both in-domain and OOD data, while requiring no OOD data during training. We also apply this framework to the "machine-aided-dialogue" corpus, a spontaneous dialogue speech task, and extend the framework in two manners. First, we introduce topic clustering which enables reliable topic confidence scores to be generated even for indistinct utterances, and second, we implement methods to effectively incorporate dialogue context. Integration of these two methods into the proposed framework significantly improves OOD detection performance, achieving a further reduction in equal error rate (EER) of 7.9 points.ja
dc.format.mimetypeapplication/pdfja
dc.language.isoengja
dc.publisherIEEEja
dc.rights© 2007 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.titleOut-of-Domain Utterance Detection Using Classification Confidences of Multiple Topicsja
dc.type.niitypeJournal Articleja
dc.identifier.ncidAA12103538ja
dc.identifier.jtitleIEEE Transactions on Audio, Speech and Language Processingja
dc.identifier.volume15ja
dc.identifier.issue1ja
dc.identifier.spage150ja
dc.identifier.epage161ja
dc.relation.doi10.1109/TASL.2006.876727ja
dc.textversionpublisherja
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。