{"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/neural-machine-translation-with-adequacy","title":"Neural Machine Translation with Adequacy-Oriented Learning","arxiv_id":"1811.08541","date":"2018-11-21","proceeding":null,"authors":["Xiang Kong","Zhaopeng Tu","Shuming Shi","Eduard Hovy","Tong Zhang"],"abstract":"Although Neural Machine Translation (NMT) models have advanced\nstate-of-the-art performance in machine translation, they face problems like\nthe inadequate translation. We attribute this to that the standard Maximum\nLikelihood Estimation (MLE) cannot judge the real translation quality due to\nits several limitations. In this work, we propose an adequacy-oriented learning\nmechanism for NMT by casting translation as a stochastic policy in\nReinforcement Learning (RL), where the reward is estimated by explicitly\nmeasuring translation adequacy. Benefiting from the sequence-level training of\nRL strategy and a more accurate reward designed specifically for translation,\nour model outperforms multiple strong baselines, including (1) standard and\ncoverage-augmented attention models with MLE-based training, and (2) advanced\nreinforcement and adversarial training strategies with rewards based on both\nword-level BLEU and character-level chrF3. Quantitative and qualitative\nanalyses on different language pairs and NMT architectures demonstrate the\neffectiveness and universality of the proposed approach.","url_abs":"http://arxiv.org/abs/1811.08541v1","url_pdf":"http://arxiv.org/pdf/1811.08541v1.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":[],"tasks":[{"task_slug":"attribute","task_name":"Attribute"},{"task_slug":"machine-translation","task_name":"Machine Translation"},{"task_slug":"nmt","task_name":"NMT"},{"task_slug":"reinforcement-learning","task_name":"Reinforcement Learning"},{"task_slug":"reinforcement-learning-1","task_name":"Reinforcement Learning (RL)"},{"task_slug":"translation","task_name":"Translation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/machine-translation-on-wmt2014-english-german","task":"Machine Translation","dataset":"WMT2014 English-German","model":"adequacy-oriented NMT","rank_in_archive_order":34,"of":91,"metrics":{"BLEU score":"28.99"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/1811.08541","atlas_url":"https://app.syntology.ai/?focus=1811.08541","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}