{"id":168476,"date":"2015-08-27T00:00:00","date_gmt":"2015-08-27T00:00:00","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/msr-research-item\/compact-lexicon-selection-with-spectral-methods\/"},"modified":"2018-10-16T20:21:14","modified_gmt":"2018-10-17T03:21:14","slug":"compact-lexicon-selection-with-spectral-methods","status":"publish","type":"msr-research-item","link":"https:\/\/www.microsoft.com\/en-us\/research\/publication\/compact-lexicon-selection-with-spectral-methods\/","title":{"rendered":"Compact Lexicon Selection with Spectral Methods"},"content":{"rendered":"<div class=\"asset-content\">\n<p>In this paper, we introduce the task of selecting compact lexicon from large, noisy gazetteers. This scenario arises often in practice, in particular spoken language understanding (SLU). We propose a simple and effective solution based on matrix decomposition techniques: canonical correlation analysis (CCA) and rank-revealing QR (RRQR) factorization. CCA is first used to derive low-dimensional gazetteer embeddings from domain-specific search logs. Then RRQR is used to find a subset of these embeddings whose span approximates the entire lexicon space. Experiments on slot tagging show that our method yields a small set of lexicon entities with average relative error reduction of > 50\\% over randomly selected lexicon.<\/p>\n<\/div>\n<p><!-- .asset-content --><\/p>\n","protected":false},"excerpt":{"rendered":"<p>In this paper, we introduce the task of selecting compact lexicon from large, noisy gazetteers. This scenario arises often in practice, in particular spoken language understanding (SLU). We propose a simple and effective solution based on matrix decomposition techniques: canonical correlation analysis (CCA) and rank-revealing QR (RRQR) factorization. CCA is first used to derive low-dimensional [&hellip;]<\/p>\n","protected":false},"featured_media":0,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","footnotes":""},"msr-content-type":[3],"msr-research-highlight":[],"research-area":[13561,13556,13545,13546],"msr-publication-type":[193716],"msr-product-type":[],"msr-focus-area":[],"msr-platform":[],"msr-download-source":[],"msr-locale":[268875],"msr-post-option":[],"msr-field-of-study":[],"msr-conference":[],"msr-journal":[],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-168476","msr-research-item","type-msr-research-item","status-publish","hentry","msr-research-area-algorithms","msr-research-area-artificial-intelligence","msr-research-area-human-language-technologies","msr-research-area-computational-sciences-mathematics","msr-locale-en_us"],"msr_publishername":"ACL - Association for Computational Linguistics","msr_edition":"Association for Computational Linguistics (ACL)","msr_affiliation":"","msr_published_date":"2015-08-27","msr_host":"","msr_duration":"","msr_version":"","msr_speaker":"","msr_other_contributors":"","msr_booktitle":"","msr_pages_string":"","msr_chapter":"","msr_isbn":"","msr_journal":"","msr_volume":"","msr_number":"","msr_editors":"","msr_series":"","msr_issue":"","msr_organization":"","msr_how_published":"","msr_notes":"","msr_highlight_text":"","msr_release_tracker_id":"","msr_original_fields_of_study":"","msr_download_urls":"","msr_external_url":"","msr_secondary_video_url":"","msr_longbiography":"","msr_microsoftintellectualproperty":1,"msr_main_download":"217039","msr_publicationurl":"","msr_doi":"","msr_publication_uploader":[{"type":"file","title":"acl15gazet_ybkim.pdf","viewUrl":"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2015\/08\/acl15gazet_ybkim.pdf","id":217039,"label_id":0}],"msr_related_uploader":"","msr_attachments":[{"id":217039,"url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2015\/08\/acl15gazet_ybkim.pdf"}],"msr-author-ordering":[{"type":"user_nicename","value":"ybkim","user_id":34985,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=ybkim"},{"type":"user_nicename","value":"rusarika","user_id":33472,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=rusarika"}],"msr_impact_theme":[],"msr_research_lab":[],"msr_event":[],"msr_group":[144736,144911,144940],"msr_project":[171150,170147],"publication":[],"video":[],"download":[],"msr_publication_type":"inproceedings","related_content":{"projects":[{"ID":171150,"post_title":"Spoken Language Understanding","post_name":"spoken-language-understanding","post_type":"msr-project","post_date":"2013-05-01 11:46:32","post_modified":"2019-08-19 14:48:51","post_status":"publish","permalink":"https:\/\/www.microsoft.com\/en-us\/research\/project\/spoken-language-understanding\/","post_excerpt":"Spoken language understanding (SLU) is an emerging field in between the areas of speech processing and natural language processing. The term spoken language understanding has largely been coined for targeted understanding of human speech directed at machines. This project covers our research on SLU tasks such as domain detection, intent determination, and slot filling, using data-driven methods. Projects Deeper Understanding: Moving\u00a0beyond shallow targeted understanding towards building domain independent SLU models. Scaling SLU: Quickly bootstrapping SLU&hellip;","_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/171150"}]}},{"ID":170147,"post_title":"Understand User's Intent from Speech and Text","post_name":"understand-users-intent-from-speech-and-text","post_type":"msr-project","post_date":"2008-12-17 11:20:26","post_modified":"2019-08-19 15:33:37","post_status":"publish","permalink":"https:\/\/www.microsoft.com\/en-us\/research\/project\/understand-users-intent-from-speech-and-text\/","post_excerpt":"Understanding what users like to do\/need to get is critical in human computer interaction. When natural user interface like speech or natural language is used in human-computer interaction, such as in a spoken dialogue system or with an internet search engine, language understanding becomes an important issue. Intent understanding is about identifying the action a user wants a computer to take or the information she\/he would like to obtain, conveyed in a spoken utterance or&hellip;","_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/170147"}]}}]},"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/168476"}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-research-item"}],"version-history":[{"count":1,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/168476\/revisions"}],"predecessor-version":[{"id":526636,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/168476\/revisions\/526636"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=168476"}],"wp:term":[{"taxonomy":"msr-content-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-content-type?post=168476"},{"taxonomy":"msr-research-highlight","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-highlight?post=168476"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=168476"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=168476"},{"taxonomy":"msr-product-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-product-type?post=168476"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=168476"},{"taxonomy":"msr-platform","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-platform?post=168476"},{"taxonomy":"msr-download-source","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-download-source?post=168476"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=168476"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=168476"},{"taxonomy":"msr-field-of-study","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-field-of-study?post=168476"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=168476"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=168476"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=168476"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=168476"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}