{"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/cross-relation-cross-bag-attention-for","title":"Cross-relation Cross-bag Attention for Distantly-supervised Relation Extraction","arxiv_id":"1812.10604","date":"2018-12-27","proceeding":null,"authors":["Yujin Yuan","Liyuan Liu","Siliang Tang","Zhongfei Zhang","Yueting Zhuang","ShiLiang Pu","Fei Wu","Xiang Ren"],"abstract":"Distant supervision leverages knowledge bases to automatically label\ninstances, thus allowing us to train relation extractor without human\nannotations. However, the generated training data typically contain massive\nnoise, and may result in poor performances with the vanilla supervised\nlearning. In this paper, we propose to conduct multi-instance learning with a\nnovel Cross-relation Cross-bag Selective Attention (C$^2$SA), which leads to\nnoise-robust training for distant supervised relation extractor. Specifically,\nwe employ the sentence-level selective attention to reduce the effect of noisy\nor mismatched sentences, while the correlation among relations were captured to\nimprove the quality of attention weights. Moreover, instead of treating all\nentity-pairs equally, we try to pay more attention to entity-pairs with a\nhigher quality. Similarly, we adopt the selective attention mechanism to\nachieve this goal. Experiments with two types of relation extractor demonstrate\nthe superiority of the proposed approach over the state-of-the-art, while\nfurther ablation studies verify our intuitions and demonstrate the\neffectiveness of our proposed two techniques.","url_abs":"http://arxiv.org/abs/1812.10604v1","url_pdf":"http://arxiv.org/pdf/1812.10604v1.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":"cross-relation-cross-bag-attention-for","repo_url":"https://github.com/yuanyu255/PCNN_C2SA","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":null,"task_name":"Relation"},{"task_slug":"relation-extraction","task_name":"Relation Extraction"},{"task_slug":"sentence","task_name":"Sentence"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1812.10604","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}