Cross-Informed Domain Adversarial Training for Noise-Robust Wake-Up Word Detection

Cited 2 time in webofscience Cited 1 time in scopus
  • Hit : 344
  • Download : 0
DC FieldValueLanguage
dc.contributor.authorLim, Hyungjunko
dc.contributor.authorKim, Younggwanko
dc.contributor.authorKim, Hoirinko
dc.date.accessioned2020-11-04T08:55:08Z-
dc.date.available2020-11-04T08:55:08Z-
dc.date.created2020-11-03-
dc.date.created2020-11-03-
dc.date.issued2020-09-
dc.identifier.citationIEEE SIGNAL PROCESSING LETTERS, v.27, pp.1769 - 1773-
dc.identifier.issn1070-9908-
dc.identifier.urihttp://hdl.handle.net/10203/277117-
dc.description.abstractA proper representation that can well express the characteristics of a word plays an important role in wake-up word detection (WWD). However, it may be easily corrupted due to various types of environmental noise occurred in the place where WWD typically works, causing unreliable performance. To deal with this practical issue, we propose a novel strategy called cross-informed domain adversarial training (CiDAT) for noise-robust WWD. In the method, additional paths were introduced to conventional domain adversarial training (DAT) to encourage its ability to generate domain-invariant representation. Experiments on the Aurora4 corpus verified that CiDAT significantly outperformed the baselines as well as conventional DAT.-
dc.languageEnglish-
dc.publisherIEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC-
dc.titleCross-Informed Domain Adversarial Training for Noise-Robust Wake-Up Word Detection-
dc.typeArticle-
dc.identifier.wosid000578457100007-
dc.identifier.scopusid2-s2.0-85094320743-
dc.type.rimsART-
dc.citation.volume27-
dc.citation.beginningpage1769-
dc.citation.endingpage1773-
dc.citation.publicationnameIEEE SIGNAL PROCESSING LETTERS-
dc.identifier.doi10.1109/LSP.2020.3026947-
dc.contributor.localauthorKim, Hoirin-
dc.contributor.nonIdAuthorKim, Younggwan-
dc.description.isOpenAccessN-
dc.type.journalArticleArticle-
dc.subject.keywordAuthorTraining-
dc.subject.keywordAuthorNoise robustness-
dc.subject.keywordAuthorEncoding-
dc.subject.keywordAuthorOptimization-
dc.subject.keywordAuthorTraining data-
dc.subject.keywordAuthorDomain adversarial training-
dc.subject.keywordAuthornoise robustness-
dc.subject.keywordAuthorwake-up word detection-
Appears in Collection
EE-Journal Papers(저널논문)
Files in This Item
There are no files associated with this item.
This item is cited by other documents in WoS
⊙ Detail Information in WoSⓡ Click to see webofscience_button
⊙ Cited 2 items in WoS Click to see citing articles in records_button

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0