Existing state of the art neural entity linking models employ attention-based bag-of-words context model and pre-trained entity embeddings bootstrapped from word embeddings to assess topic level context compatibility. However, the latent entity type information in the immediate context of the mention is neglected, which causes the models often link mentions to incorrect entities with incorrect type. To tackle this problem, we propose to inject latent entity type information into the entity embeddings based on pre-trained BERT. In addition, we integrate a BERT-based entity similarity score into the local context model of a state-of-the-art model to better capture latent entity type information. Our model significantly outperforms the state-of-the-art entity linking models on standard benchmark (AIDA-CoNLL). Detailed experiment analysis demonstrates that our model corrects most of the type errors produced by the direct baseline.<\/p>\n","protected":false},"excerpt":{"rendered":"
Existing state of the art neural entity linking models employ attention-based bag-of-words context model and pre-trained entity embeddings bootstrapped from word embeddings to assess topic level context compatibility. However, the latent entity type information in the immediate context of the mention is neglected, which causes the models often link mentions to incorrect entities with incorrect […]<\/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":"","msr-author-ordering":null,"msr_publishername":"Association for the Advancement of Artificial Intelligence (AAAI)","msr_publisher_other":"","msr_booktitle":"","msr_chapter":"","msr_edition":"","msr_editors":"","msr_how_published":"","msr_isbn":"","msr_issue":"","msr_journal":"","msr_number":"","msr_organization":"","msr_pages_string":"","msr_page_range_start":"7529","msr_page_range_end":"7537","msr_series":"","msr_volume":"","msr_copyright":"","msr_conference_name":"National Conference on Artificial Intelligence","msr_doi":"","msr_arxiv_id":"","msr_s2_paper_id":"","msr_mag_id":"2997110185","msr_pubmed_id":"","msr_other_authors":"","msr_other_contributors":"","msr_speaker":"","msr_award":"","msr_affiliation":"","msr_institution":"","msr_host":"","msr_version":"","msr_duration":"","msr_original_fields_of_study":"","msr_release_tracker_id":"","msr_s2_match_type":"","msr_citation_count_updated":"","msr_published_date":"2020-4-3","msr_highlight_text":"","msr_notes":"","msr_longbiography":"","msr_publicationurl":"","msr_external_url":"","msr_secondary_video_url":"","msr_conference_url":"","msr_journal_url":"","msr_s2_pdf_url":"","msr_year":0,"msr_citation_count":0,"msr_influential_citations":0,"msr_reference_count":0,"msr_s2_match_confidence":0,"msr_microsoftintellectualproperty":true,"msr_s2_open_access":false,"msr_s2_author_ids":[],"msr_pub_ids":[],"msr_hide_image_in_river":0,"footnotes":""},"msr-research-highlight":[],"research-area":[13556,13545],"msr-publication-type":[193716],"msr-publisher":[],"msr-focus-area":[],"msr-locale":[268875],"msr-post-option":[],"msr-field-of-study":[246694,248794,246691,248791,248788,248797,246808],"msr-conference":[],"msr-journal":[],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-714409","msr-research-item","type-msr-research-item","status-publish","hentry","msr-research-area-artificial-intelligence","msr-research-area-human-language-technologies","msr-locale-en_us","msr-field-of-study-artificial-intelligence","msr-field-of-study-bootstrapping","msr-field-of-study-computer-science","msr-field-of-study-context-model","msr-field-of-study-entity-linking","msr-field-of-study-entity-type","msr-field-of-study-natural-language-processing"],"msr_publishername":"Association for the Advancement of Artificial Intelligence (AAAI)","msr_edition":"","msr_affiliation":"","msr_published_date":"2020-4-3","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":"","msr_publicationurl":"","msr_doi":"","msr_publication_uploader":[{"type":"doi","viewUrl":"false","id":"false","title":"10.1609\/AAAI.V34I05.6251","label_id":"243106","label":0},{"type":"url","viewUrl":"false","id":"false","title":"https:\/\/aaai.org\/ojs\/index.php\/AAAI\/article\/view\/6251","label_id":"243109","label":0},{"type":"url","viewUrl":"false","id":"false","title":"https:\/\/dblp.uni-trier.de\/db\/conf\/aaai\/aaai2020.html#ChenWJL20","label_id":"243109","label":0},{"type":"url","viewUrl":"false","id":"false","title":"https:\/\/ojs.aaai.org\/index.php\/AAAI\/article\/view\/6251","label_id":"243109","label":0}],"msr_related_uploader":"","msr_citation_count":0,"msr_citation_count_updated":"","msr_s2_paper_id":"","msr_influential_citations":0,"msr_reference_count":0,"msr_arxiv_id":"","msr_s2_author_ids":[],"msr_s2_open_access":false,"msr_s2_pdf_url":null,"msr_attachments":[],"msr-author-ordering":[{"type":"text","value":"Shuang Chen","user_id":0,"rest_url":false},{"type":"text","value":"Jinpeng Wang","user_id":0,"rest_url":false},{"type":"text","value":"Feng Jiang","user_id":0,"rest_url":false},{"type":"user_nicename","value":"Chin-Yew Lin","user_id":31493,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Chin-Yew Lin"}],"msr_impact_theme":[],"msr_research_lab":[199560],"msr_event":[],"msr_group":[144919],"msr_project":[792599,717721,714646],"publication":[],"video":[],"msr-tool":[],"msr_publication_type":"inproceedings","related_content":{"projects":[{"ID":792599,"post_title":"Table Interpretation","post_name":"table-interpretation","post_type":"msr-project","post_date":"2021-11-05 02:02:36","post_modified":"2024-09-25 11:42:48","post_status":"publish","permalink":"https:\/\/www.microsoft.com\/en-us\/research\/project\/table-interpretation\/","post_excerpt":"Bringing out the power of semantics in tabular data Tables are commonly used to organize information, playing a key role in data analytics, scientific research, and business communication. The ability to automatically extract semantics in tables can empower many downstream applications such as data analytics, robotic process automation (RPA), knowledge base population, etc. In this project, we explore multiple aspects of semantic table understanding and real-world applications of such technologies. One of the outcomes of…","_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/792599"}]}},{"ID":717721,"post_title":"ezPitch: Connecting Salespersons and Customers through Relevant News","post_name":"ezpitch-connecting-salespersons-and-customers-through-relevant-news","post_type":"msr-project","post_date":"2021-01-19 08:25:06","post_modified":"2021-01-19 08:28:44","post_status":"publish","permalink":"https:\/\/www.microsoft.com\/en-us\/research\/project\/ezpitch-connecting-salespersons-and-customers-through-relevant-news\/","post_excerpt":"The goal of ezPitch is connecting salespersons and customers through relevant news. Why is this important? In the daily work, the sales persons need to search, track and explore the related news about customers before talking to them. For example, if there is management change in the customer\u2019s company. The sales person may need to find a way to re-build the relationship with the new leadership. If the customer\u2019s company announces an earnings report which…","_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/717721"}]}},{"ID":714646,"post_title":"VERT: Versatile Entity Recognition & Disambiguation Toolkit","post_name":"vert-versatile-entity-recognition-disambiguation-toolkit","post_type":"msr-project","post_date":"2020-12-30 02:54:35","post_modified":"2021-10-13 21:15:01","post_status":"publish","permalink":"https:\/\/www.microsoft.com\/en-us\/research\/project\/vert-versatile-entity-recognition-disambiguation-toolkit\/","post_excerpt":"While knowledge about entities is a key building block in the mentioned systems, creating effective\/efficient models for real-world scenarios remains a challenge (tech\/data\/real workloads). Based on such needs, we've created VERT - a Versatile Entity Recognition & Disambiguation Toolkit. VERT is a pragmatic toolkit that combines rules and ML, offering both powerful pretrained models for core entity types (recognition and linking) and the easy creation of custom models. Custom models use our deep learning-based NER\/EL…","_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/714646"}]}}]},"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/714409","targetHints":{"allow":["GET"]}}],"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\/714409\/revisions"}],"predecessor-version":[{"id":714412,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/714409\/revisions\/714412"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=714409"}],"wp:term":[{"taxonomy":"msr-research-highlight","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-highlight?post=714409"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=714409"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=714409"},{"taxonomy":"msr-publisher","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-publisher?post=714409"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=714409"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=714409"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=714409"},{"taxonomy":"msr-field-of-study","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-field-of-study?post=714409"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=714409"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=714409"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=714409"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=714409"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}