{"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/semantic-specialisation-of-distributional","title":"Semantic Specialisation of Distributional Word Vector Spaces using Monolingual and Cross-Lingual Constraints","arxiv_id":"1706.00374","date":"2017-06-01","proceeding":null,"authors":["Nikola Mrkšić","Ivan Vulić","Diarmuid Ó Séaghdha","Ira Leviant","Roi Reichart","Milica Gašić","Anna Korhonen","Steve Young"],"abstract":"We present Attract-Repel, an algorithm for improving the semantic quality of\nword vectors by injecting constraints extracted from lexical resources.\nAttract-Repel facilitates the use of constraints from mono- and cross-lingual\nresources, yielding semantically specialised cross-lingual vector spaces. Our\nevaluation shows that the method can make use of existing cross-lingual\nlexicons to construct high-quality vector spaces for a plethora of different\nlanguages, facilitating semantic transfer from high- to lower-resource ones.\nThe effectiveness of our approach is demonstrated with state-of-the-art results\non semantic similarity datasets in six languages. We next show that\nAttract-Repel-specialised vectors boost performance in the downstream task of\ndialogue state tracking (DST) across multiple languages. Finally, we show that\ncross-lingual vector spaces produced by our algorithm facilitate the training\nof multilingual DST models, which brings further performance improvements.","url_abs":"http://arxiv.org/abs/1706.00374v1","url_pdf":"http://arxiv.org/pdf/1706.00374v1.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":"semantic-specialisation-of-distributional","repo_url":"https://github.com/nmrksic/attract-repel","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"semantic-specialisation-of-distributional","repo_url":"https://github.com/zliucr/mixed-language-training","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"dialogue-state-tracking","task_name":"Dialogue State Tracking"},{"task_slug":"semantic-similarity","task_name":"Semantic Similarity"},{"task_slug":"semantic-textual-similarity","task_name":"Semantic Textual Similarity"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1706.00374","atlas_url":"https://app.syntology.ai/?focus=1706.00374","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}