{"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/take-and-took-gaggle-and-goose-book-and-read","title":"Take and Took, Gaggle and Goose, Book and Read: Evaluating the Utility of Vector Differences for Lexical Relation Learning","arxiv_id":"1509.01692","date":"2015-09-05","proceeding":"ACL 2016 8","authors":["Ekaterina Vylomova","Laura Rimell","Trevor Cohn","Timothy Baldwin"],"abstract":"Recent work on word embeddings has shown that simple vector subtraction over\npre-trained embeddings is surprisingly effective at capturing different lexical\nrelations, despite lacking explicit supervision. Prior work has evaluated this\nintriguing result using a word analogy prediction formulation and hand-selected\nrelations, but the generality of the finding over a broader range of lexical\nrelation types and different learning settings has not been evaluated. In this\npaper, we carry out such an evaluation in two learning settings: (1) spectral\nclustering to induce word relations, and (2) supervised learning to classify\nvector differences into relation types. We find that word embeddings capture a\nsurprising amount of information, and that, under suitable supervised training,\nvector subtraction generalises well to a broad range of relations, including\nover unseen lexical items.","url_abs":"http://arxiv.org/abs/1509.01692v4","url_pdf":"http://arxiv.org/pdf/1509.01692v4.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":"take-and-took-gaggle-and-goose-book-and-read","repo_url":"https://github.com/ivri/DiffVec","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"clustering","task_name":"Clustering"},{"task_slug":null,"task_name":"Relation"},{"task_slug":"word-embeddings","task_name":"Word Embeddings"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1509.01692","atlas_url":"https://app.syntology.ai/?focus=1509.01692","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}