{"id":473901,"date":"2018-03-16T00:44:50","date_gmt":"2018-03-16T07:44:50","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/?post_type=msr-project&#038;p=473901"},"modified":"2018-03-16T02:30:28","modified_gmt":"2018-03-16T09:30:28","slug":"question-generation-qg","status":"publish","type":"msr-project","link":"https:\/\/www.microsoft.com\/en-us\/research\/project\/question-generation-qg\/","title":{"rendered":"Question Generation (QG)"},"content":{"rendered":"<p><strong>Question Generation (QG)<\/strong> aims to generate natural language questions based on given contents (knowledge base triples, tables, sentences, or images), where the generated questions need to be able to be\u00a0answered by the contents. The motivation of QG task is two-fold: (i) transforming customized contents into Q-A pairs, which can\u00a0be easily\u00a0used to build customized\u00a0QA or dialogue systems; (ii) generating large scale Q-A pairs with acceptable quality, which can either be used as additional QA model training data, or improve the efficiency of human annotation on QA dataset construction.<\/p>\n<p>Recently, we are working on three\u00a0QG tasks, including\u00a0<span style=\"color: #0000ff\">Structured data-based QG<\/span>, which geneartes questions from knowledge base sub-graphs or\u00a0semi-structured tables,\u00a0<span style=\"color: #0000ff\">Text-based QG<\/span>, which generates questions from natural language texts,\u00a0and <span style=\"color: #0000ff\">Image-based QG<\/span>, which generates questions from images. Some\u00a0of\u00a0our work\u00a0are listed\u00a0below:<\/p>\n<ul>\n<li>Yikang Li, Nan Duan<span style=\"color: #000000;text-indent: 0px;letter-spacing: normal;font-family: 'Lucida Grande', helvetica, arial, verdana, sans-serif;font-size: medium;font-style: normal;font-weight: 400;float: none;background-color: #ffffff\">,\u00a0<\/span>Bolei Zhou<span style=\"color: #000000;text-indent: 0px;letter-spacing: normal;font-family: 'Lucida Grande', helvetica, arial, verdana, sans-serif;font-size: medium;font-style: normal;font-weight: 400;float: none;background-color: #ffffff\">,\u00a0<\/span>Xiao Chu<span style=\"color: #000000;text-indent: 0px;letter-spacing: normal;font-family: 'Lucida Grande', helvetica, arial, verdana, sans-serif;font-size: medium;font-style: normal;font-weight: 400;float: none;background-color: #ffffff\">,\u00a0<\/span>Wanli Ouyang<span style=\"color: #000000;text-indent: 0px;letter-spacing: normal;font-family: 'Lucida Grande', helvetica, arial, verdana, sans-serif;font-size: medium;font-style: normal;font-weight: 400;float: none;background-color: #ffffff\">,\u00a0<\/span>Xiaogang Wang, \u201c<strong>Visual Question Generation as Dual Task of Visual Question Answering<\/strong>\u201d,\u00a0CVPR, 2018. <span style=\"color: #0000ff\">(Image-based QG)<\/span><\/li>\n<li>Duyu Tang, Nan Duan, Zhao Yan, Zhirui Zhang, Yibo Sun, Shujie Liu, Yuanhua Lv, Ming Zhou, \u201c<strong>Learning to Collaborate for Question Answering and Asking<\/strong>\u201d, NAACL, 2018. <span style=\"color: #0000ff\">(Structured data-based QG)<\/span><\/li>\n<li>Junwei Bao, Duyu Tang, Nan Duan, Zhao Yan, Yuanhua Lv, Ming Zhou, Tiejun Zhao, \u201c<strong>Table-to-Text: Describing Table Region with Natural Language<\/strong>\u201d, AAAI, 2018. <span style=\"color: #0000ff\">(Structured data-based QG)<\/span><\/li>\n<li>Nan Duan, Duyu Tang, Peng Chen, Ming Zhou,\u00a0\u201c<strong>Question Generation for Question Answering<\/strong>\u201d,\u00a0EMNLP, 2017. <span style=\"color: #0000ff\">(Text-based QG)<\/span><\/li>\n<li>Duyu Tang, Nan Duan<span style=\"color: #000000;text-indent: 0px;letter-spacing: normal;font-family: 'Lucida Grande', helvetica, arial, verdana, sans-serif;font-size: medium;font-style: normal;font-weight: 400;float: none;background-color: #ffffff\">,\u00a0<\/span>Tao Qin<span style=\"color: #000000;text-indent: 0px;letter-spacing: normal;font-family: 'Lucida Grande', helvetica, arial, verdana, sans-serif;font-size: medium;font-style: normal;font-weight: 400;float: none;background-color: #ffffff\">,\u00a0<\/span>Zhao Yan<span style=\"color: #000000;text-indent: 0px;letter-spacing: normal;font-family: 'Lucida Grande', helvetica, arial, verdana, sans-serif;font-size: medium;font-style: normal;font-weight: 400;float: none;background-color: #ffffff\">,\u00a0<\/span>Ming Zhou, \u201c<strong>Question Answering and Question Generation as Dual Tasks<\/strong>\u201d, arXiv, 2017. <span style=\"color: #0000ff\">(Text-based QG)<\/span><\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>Question Generation (QG) aims to generate natural language questions based on given contents (knowledge base triples, tables, sentences, or images), where the generated questions need to be able to be\u00a0answered by the contents. The motivation of QG task is two-fold: (i) transforming customized contents into Q-A pairs, which can\u00a0be easily\u00a0used to build customized\u00a0QA or dialogue [&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":""},"research-area":[13545],"msr-locale":[268875],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-473901","msr-project","type-msr-project","status-publish","hentry","msr-research-area-human-language-technologies","msr-locale-en_us","msr-archive-status-active"],"msr_project_start":"","related-publications":[],"related-downloads":[],"related-videos":[],"related-groups":[],"related-events":[],"related-opportunities":[],"related-posts":[],"related-articles":[],"tab-content":[],"slides":[],"related-researchers":[{"type":"user_nicename","display_name":"Yaobo Liang","user_id":36036,"people_section":"Section name 1","alias":"yalia"}],"msr_research_lab":[199560],"msr_impact_theme":[],"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/473901","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-project"}],"version-history":[{"count":13,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/473901\/revisions"}],"predecessor-version":[{"id":473946,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/473901\/revisions\/473946"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=473901"}],"wp:term":[{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=473901"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=473901"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=473901"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=473901"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}