{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/indirect-supervision-for-relation-extraction","title":"Indirect Supervision for Relation Extraction using Question-Answer Pairs","arxiv_id":"1710.11169","date":"2017-10-30","proceeding":null,"authors":["Zeqiu Wu","Xiang Ren","Frank F. Xu","Ji Li","Jiawei Han"],"abstract":"Automatic relation extraction (RE) for types of interest is of great\nimportance for interpreting massive text corpora in an efficient manner.\nTraditional RE models have heavily relied on human-annotated corpus for\ntraining, which can be costly in generating labeled data and become obstacles\nwhen dealing with more relation types. Thus, more RE extraction systems have\nshifted to be built upon training data automatically acquired by linking to\nknowledge bases (distant supervision). However, due to the incompleteness of\nknowledge bases and the context-agnostic labeling, the training data collected\nvia distant supervision (DS) can be very noisy. In recent years, as increasing\nattention has been brought to tackling question-answering (QA) tasks, user\nfeedback or datasets of such tasks become more accessible. In this paper, we\npropose a novel framework, ReQuest, to leverage question-answer pairs as an\nindirect source of supervision for relation extraction, and study how to use\nsuch supervision to reduce noise induced from DS. Our model jointly embeds\nrelation mentions, types, QA entity mention pairs and text features in two\nlow-dimensional spaces (RE and QA), where objects with same relation types or\nsemantically similar question-answer pairs have similar representations. Shared\nfeatures connect these two spaces, carrying clearer semantic knowledge from\nboth sources. ReQuest, then use these learned embeddings to estimate the types\nof test relation mentions. We formulate a global objective function and adopt a\nnovel margin-based QA loss to reduce noise in DS by exploiting semantic\nevidence from the QA dataset. Our experimental results achieve an average of\n11% improvement in F1 score on two public RE datasets combined with TREC QA\ndataset.","url_abs":"http://arxiv.org/abs/1710.11169v2","url_pdf":"http://arxiv.org/pdf/1710.11169v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"indirect-supervision-for-relation-extraction","repo_url":"https://github.com/ellenmellon/ReQuest","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"indirect-supervision-for-relation-extraction","repo_url":"https://github.com/shanzhenren/CoType","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"indirect-supervision-for-relation-extraction","repo_url":"https://github.com/ink-usc/request","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"question-answering","task_name":"Question Answering"},{"task_slug":null,"task_name":"Relation"},{"task_slug":"relation-extraction","task_name":"Relation Extraction"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}