{"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/pair2vec-compositional-word-pair-embeddings","title":"pair2vec: Compositional Word-Pair Embeddings for Cross-Sentence Inference","arxiv_id":"1810.08854","date":"2018-10-20","proceeding":"NAACL 2019 6","authors":["Mandar Joshi","Eunsol Choi","Omer Levy","Daniel S. Weld","Luke Zettlemoyer"],"abstract":"Reasoning about implied relationships (e.g., paraphrastic, common sense,\nencyclopedic) between pairs of words is crucial for many cross-sentence\ninference problems. This paper proposes new methods for learning and using\nembeddings of word pairs that implicitly represent background knowledge about\nsuch relationships. Our pairwise embeddings are computed as a compositional\nfunction on word representations, which is learned by maximizing the pointwise\nmutual information (PMI) with the contexts in which the two words co-occur. We\nadd these representations to the cross-sentence attention layer of existing\ninference models (e.g. BiDAF for QA, ESIM for NLI), instead of extending or\nreplacing existing word embeddings. Experiments show a gain of 2.7% on the\nrecently released SQuAD2.0 and 1.3% on MultiNLI. Our representations also aid\nin better generalization with gains of around 6-7% on adversarial SQuAD\ndatasets, and 8.8% on the adversarial entailment test set by Glockner et al.\n(2018).","url_abs":"http://arxiv.org/abs/1810.08854v2","url_pdf":"http://arxiv.org/pdf/1810.08854v2.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":"pair2vec-compositional-word-pair-embeddings","repo_url":"https://github.com/mandarjoshi90/pair2vec","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"pair2vec-compositional-word-pair-embeddings","repo_url":"https://github.com/asahi417/relbert","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"pair2vec-compositional-word-pair-embeddings","repo_url":"https://github.com/ghlee0304/NLP-with-tensorflow-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"common-sense-reasoning","task_name":"Common Sense Reasoning"},{"task_slug":"sentence","task_name":"Sentence"},{"task_slug":"word-embeddings","task_name":"Word Embeddings"}],"methods":[{"method_slug":"esim","method_name":"ESIM"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1810.08854","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}