{"links":{},"id":82944,"metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00082944","sets":["1164:5159:6679:6825"]},"path":["6825"],"owner":"11","recid":"82944","title":["Comparison of Discriminative Models for Lexicon Optimization for ASR of Agglutinative Language"],"pubdate":{"attribute_name":"公開日","attribute_value":"2012-07-12"},"_buckets":{"deposit":"282b839d-c3e4-401c-a1ea-337cf7dc0b6f"},"_deposit":{"id":"82944","pid":{"type":"depid","value":"82944","revision_id":0},"owners":[11],"status":"published","created_by":11},"item_title":"Comparison of Discriminative Models for Lexicon Optimization for ASR of Agglutinative Language","author_link":["0","0"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"Comparison of Discriminative Models for Lexicon Optimization for ASR of Agglutinative Language"},{"subitem_title":"Comparison of Discriminative Models for Lexicon Optimization for ASR of Agglutinative Language","subitem_title_language":"en"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"高精度音声認識","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2012-07-12","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"School of Informatics, Kyoto University"},{"subitem_text_value":"School of Informatics, Kyoto University"},{"subitem_text_value":"Institute of Information Engineering, Xinjiang University"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"School of Informatics, Kyoto University","subitem_text_language":"en"},{"subitem_text_value":"School of Informatics, Kyoto University","subitem_text_language":"en"},{"subitem_text_value":"Institute of Information Engineering, Xinjiang University","subitem_text_language":"en"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"eng"}]},"item_publisher":{"attribute_name":"出版者","attribute_value_mlt":[{"subitem_publisher":"情報処理学会","subitem_publisher_language":"ja"}]},"publish_status":"0","weko_shared_id":-1,"item_file_price":{"attribute_name":"Billing file","attribute_type":"file","attribute_value_mlt":[{"url":{"url":"https://ipsj.ixsq.nii.ac.jp/record/82944/files/IPSJ-SLP12092013.pdf"},"date":[{"dateType":"Available","dateValue":"2014-07-12"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-SLP12092013.pdf","filesize":[{"value":"273.3 kB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"660","billingrole":"5"},{"tax":["include_tax"],"price":"330","billingrole":"6"},{"tax":["include_tax"],"price":"0","billingrole":"22"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"edfb6e90-b680-4958-b47f-c1d3578ff9c9","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2012 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Mijit, Ablimit"},{"creatorName":"Tatsuya, Kawahara"},{"creatorName":"Askar, Hamdulla"}],"nameIdentifiers":[{}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Mijit, Ablimit","creatorNameLang":"en"},{"creatorName":"Tatsuya, Kawahara","creatorNameLang":"en"},{"creatorName":"Askar, Hamdulla","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AN10442647","subitem_source_identifier_type":"NCID"}]},"item_4_textarea_12":{"attribute_name":"Notice","attribute_value_mlt":[{"subitem_textarea_value":"SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc."}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_18gh","resourcetype":"technical report"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"For automatic speech recognition (ASR) of agglutinative languages, selection of lexical unit is not obvious. Morpheme unit is usually adopted to ensure the sufficient coverage, but many morphemes are short, resulting in weak constraints and possible confusions. We have proposed a discriminative approach to select lexical entries which will directly contribute to ASR error reduction, considering not only linguistic constraint but also acoustic-phonetic confusability. It is based on an evaluation function for each word defined by a set of features and their weights, which are optimized by the difference of word error rates (WERs) by the morpheme-based model and those by the word-based model. In this paper, we investigate several discriminative models to realize this scheme. Specifically, we implement with Support Vector Machines (SVM) and Logistic Regression (LR) model as well as simple perceptron. Experimental evaluations on Uyghur LVCSR show that SVM and LR are more robustly trained and SVM results in the best performance with a large dimension of features.","subitem_description_type":"Other"}]},"item_4_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"For automatic speech recognition (ASR) of agglutinative languages, selection of lexical unit is not obvious. Morpheme unit is usually adopted to ensure the sufficient coverage, but many morphemes are short, resulting in weak constraints and possible confusions. We have proposed a discriminative approach to select lexical entries which will directly contribute to ASR error reduction, considering not only linguistic constraint but also acoustic-phonetic confusability. It is based on an evaluation function for each word defined by a set of features and their weights, which are optimized by the difference of word error rates (WERs) by the morpheme-based model and those by the word-based model. In this paper, we investigate several discriminative models to realize this scheme. Specifically, we implement with Support Vector Machines (SVM) and Logistic Regression (LR) model as well as simple perceptron. Experimental evaluations on Uyghur LVCSR show that SVM and LR are more robustly trained and SVM results in the best performance with a large dimension of features.","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"4","bibliographic_titles":[{"bibliographic_title":"研究報告音声言語情報処理（SLP）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2012-07-12","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"13","bibliographicVolumeNumber":"2012-SLP-92"}]},"relation_version_is_last":true,"weko_creator_id":"11"},"created":"2025-01-18T23:36:39.019316+00:00","updated":"2025-01-21T18:46:57.396826+00:00"}