The growing demand for structured\u00a0knowledge has led to great interest in\u00a0relation extraction, especially in cases\u00a0with limited supervision. However,\u00a0existing distance supervision approaches\u00a0only extract relations expressed in single\u00a0sentences. In general, cross-sentence\u00a0relation extraction is under-explored, even\u00a0in the supervised-learning setting. In this\u00a0paper, we propose the first approach for\u00a0applying distant supervision to crosssentence\u00a0relation extraction. At the core\u00a0of our approach is a graph representation\u00a0that can incorporate both standard\u00a0dependencies and discourse relations,\u00a0thus providing a unifying way to model\u00a0relations within and across sentences. We\u00a0extract features from multiple paths in this\u00a0graph, increasing accuracy and robustness\u00a0when confronted with linguistic variation\u00a0and analysis error. Experiments on an\u00a0important extraction task for precision\u00a0medicine show that our approach can learn\u00a0an accurate cross-sentence extractor, using\u00a0only a small existing knowledge base and\u00a0unlabeled text from biomedical research\u00a0articles. Compared to the existing distant\u00a0supervision paradigm, our approach\u00a0extracted twice as many relations at\u00a0similar precision, thus demonstrating the\u00a0prevalence of cross-sentence relations and\u00a0the promise of our approach.<\/p>\n","protected":false},"excerpt":{"rendered":"
The growing demand for structured\u00a0knowledge has led to great interest in\u00a0relation extraction, especially in cases\u00a0with limited supervision. However,\u00a0existing distance supervision approaches\u00a0only extract relations expressed in single\u00a0sentences. In general, cross-sentence\u00a0relation extraction is under-explored, even\u00a0in the supervised-learning setting. In this\u00a0paper, we propose the first approach for\u00a0applying distant supervision to crosssentence\u00a0relation extraction. At the core\u00a0of our approach is […]<\/p>\n","protected":false},"featured_media":0,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"footnotes":""},"msr-content-type":[3],"msr-research-highlight":[],"research-area":[13553],"msr-publication-type":[193716],"msr-product-type":[],"msr-focus-area":[],"msr-platform":[],"msr-download-source":[],"msr-locale":[268875],"msr-field-of-study":[],"msr-conference":[],"msr-journal":[],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-394730","msr-research-item","type-msr-research-item","status-publish","hentry","msr-research-area-medical-health-genomics","msr-locale-en_us"],"msr_publishername":"","msr_edition":"Fifteenth Conference of the European Association for Computational Linguistics (EACL), 2017","msr_affiliation":"","msr_published_date":"2016-09-30","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":"386933","msr_publicationurl":"","msr_doi":"","msr_publication_uploader":[{"type":"file","title":"eacl17","viewUrl":"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2017\/05\/eacl17.pdf","id":386933,"label_id":0}],"msr_related_uploader":"","msr_attachments":[],"msr-author-ordering":[{"type":"user_nicename","value":"chrisq","user_id":31430,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=chrisq"},{"type":"user_nicename","value":"hoifung","user_id":32016,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=hoifung"}],"msr_impact_theme":[],"msr_research_lab":[],"msr_event":[],"msr_group":[952050],"msr_project":[393227],"publication":[],"video":[],"download":[],"msr_publication_type":"inproceedings","_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/394730"}],"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":4,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/394730\/revisions"}],"predecessor-version":[{"id":417269,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/394730\/revisions\/417269"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=394730"}],"wp:term":[{"taxonomy":"msr-content-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-content-type?post=394730"},{"taxonomy":"msr-research-highlight","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-highlight?post=394730"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=394730"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=394730"},{"taxonomy":"msr-product-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-product-type?post=394730"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=394730"},{"taxonomy":"msr-platform","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-platform?post=394730"},{"taxonomy":"msr-download-source","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-download-source?post=394730"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=394730"},{"taxonomy":"msr-field-of-study","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-field-of-study?post=394730"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=394730"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=394730"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=394730"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=394730"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}