{"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/multilingual-neural-machine-translation-with-2","title":"Multilingual Neural Machine Translation with Knowledge Distillation","arxiv_id":"1902.10461","date":"2019-02-27","proceeding":"ICLR 2019 5","authors":["Xu Tan","Yi Ren","Di He","Tao Qin","Zhou Zhao","Tie-Yan Liu"],"abstract":"Multilingual machine translation, which translates multiple languages with a\nsingle model, has attracted much attention due to its efficiency of offline\ntraining and online serving. However, traditional multilingual translation\nusually yields inferior accuracy compared with the counterpart using individual\nmodels for each language pair, due to language diversity and model capacity\nlimitations. In this paper, we propose a distillation-based approach to boost\nthe accuracy of multilingual machine translation. Specifically, individual\nmodels are first trained and regarded as teachers, and then the multilingual\nmodel is trained to fit the training data and match the outputs of individual\nmodels simultaneously through knowledge distillation. Experiments on IWSLT, WMT\nand Ted talk translation datasets demonstrate the effectiveness of our method.\nParticularly, we show that one model is enough to handle multiple languages (up\nto 44 languages in our experiment), with comparable or even better accuracy\nthan individual models.","url_abs":"http://arxiv.org/abs/1902.10461v3","url_pdf":"http://arxiv.org/pdf/1902.10461v3.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":"multilingual-neural-machine-translation-with-2","repo_url":"https://github.com/RayeRen/multilingual-kd-pytorch","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}}],"tasks":[{"task_slug":"diversity","task_name":"Diversity"},{"task_slug":"knowledge-distillation","task_name":"Knowledge Distillation"},{"task_slug":"machine-translation","task_name":"Machine Translation"},{"task_slug":"translation","task_name":"Translation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1902.10461","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}