{"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/dual-learning-for-machine-translation","title":"Dual Learning for Machine Translation","arxiv_id":"1611.00179","date":"2016-11-01","proceeding":"NeurIPS 2016 12","authors":["Yingce Xia","Di He","Tao Qin","Li-Wei Wang","Nenghai Yu","Tie-Yan Liu","Wei-Ying Ma"],"abstract":"While neural machine translation (NMT) is making good progress in the past\ntwo years, tens of millions of bilingual sentence pairs are needed for its\ntraining. However, human labeling is very costly. To tackle this training data\nbottleneck, we develop a dual-learning mechanism, which can enable an NMT\nsystem to automatically learn from unlabeled data through a dual-learning game.\nThis mechanism is inspired by the following observation: any machine\ntranslation task has a dual task, e.g., English-to-French translation (primal)\nversus French-to-English translation (dual); the primal and dual tasks can form\na closed loop, and generate informative feedback signals to train the\ntranslation models, even if without the involvement of a human labeler. In the\ndual-learning mechanism, we use one agent to represent the model for the primal\ntask and the other agent to represent the model for the dual task, then ask\nthem to teach each other through a reinforcement learning process. Based on the\nfeedback signals generated during this process (e.g., the language-model\nlikelihood of the output of a model, and the reconstruction error of the\noriginal sentence after the primal and dual translations), we can iteratively\nupdate the two models until convergence (e.g., using the policy gradient\nmethods). We call the corresponding approach to neural machine translation\n\\emph{dual-NMT}. Experiments show that dual-NMT works very well on\nEnglish$\\leftrightarrow$French translation; especially, by learning from\nmonolingual data (with 10% bilingual data for warm start), it achieves a\ncomparable accuracy to NMT trained from the full bilingual data for the\nFrench-to-English translation task.","url_abs":"http://arxiv.org/abs/1611.00179v1","url_pdf":"http://arxiv.org/pdf/1611.00179v1.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":"dual-learning-for-machine-translation","repo_url":"https://github.com/NonameAuPlatal/Dual_Learning","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"language-modeling","task_name":"Language Modeling"},{"task_slug":"language-modelling","task_name":"Language Modelling"},{"task_slug":"machine-translation","task_name":"Machine Translation"},{"task_slug":"nmt","task_name":"NMT"},{"task_slug":"policy-gradient-methods","task_name":"Policy Gradient Methods"},{"task_slug":"reinforcement-learning","task_name":"Reinforcement Learning"},{"task_slug":"sentence","task_name":"Sentence"},{"task_slug":"translation","task_name":"Translation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1611.00179","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}