{"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/improving-neural-machine-translation-models","title":"Improving Neural Machine Translation Models with Monolingual Data","arxiv_id":"1511.06709","date":"2015-11-20","proceeding":"ACL 2016 8","authors":["Rico Sennrich","Barry Haddow","Alexandra Birch"],"abstract":"Neural Machine Translation (NMT) has obtained state-of-the art performance\nfor several language pairs, while only using parallel data for training.\nTarget-side monolingual data plays an important role in boosting fluency for\nphrase-based statistical machine translation, and we investigate the use of\nmonolingual data for NMT. In contrast to previous work, which combines NMT\nmodels with separately trained language models, we note that encoder-decoder\nNMT architectures already have the capacity to learn the same information as a\nlanguage model, and we explore strategies to train with monolingual data\nwithout changing the neural network architecture. By pairing monolingual\ntraining data with an automatic back-translation, we can treat it as additional\nparallel training data, and we obtain substantial improvements on the WMT 15\ntask English<->German (+2.8-3.7 BLEU), and for the low-resourced IWSLT 14 task\nTurkish->English (+2.1-3.4 BLEU), obtaining new state-of-the-art results. We\nalso show that fine-tuning on in-domain monolingual and parallel data gives\nsubstantial improvements for the IWSLT 15 task English->German.","url_abs":"http://arxiv.org/abs/1511.06709v4","url_pdf":"http://arxiv.org/pdf/1511.06709v4.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":"improving-neural-machine-translation-models","repo_url":"https://github.com/josephch405/curriculum-nmt","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"improving-neural-machine-translation-models","repo_url":"https://github.com/surafelml/Afro-NMT","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}}],"tasks":[{"task_slug":"cross-lingual-bitext-mining","task_name":"Cross-Lingual Bitext Mining"},{"task_slug":"decoder","task_name":"Decoder"},{"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":"translation","task_name":"Translation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/cross-lingual-bitext-mining-on-bucc-french-to","task":"Cross-Lingual Bitext Mining","dataset":"BUCC French-to-English","model":"Monolingual training data","rank_in_archive_order":3,"of":3,"metrics":{"F1 score":"75.8"},"uses_additional_data":false},{"leaderboard":"/sota/cross-lingual-bitext-mining-on-bucc-german-to","task":"Cross-Lingual Bitext Mining","dataset":"BUCC German-to-English","model":"Monolingual training data","rank_in_archive_order":3,"of":3,"metrics":{"F1 score":"76.9"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1511.06709","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1511.06709"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/surafelml/Afro-NMT","reach":{"status":"ok"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/josephch405/curriculum-nmt","reach":null}],"summary":{"ran_honours":3,"ran_draft_wrong":3},"by_repo_kind":{"listed":{"samples":6,"ran":6,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"e40cffce14ff31fa","entry":"get_difficulty_scores","repo":"josephch405/curriculum-nmt","repo_kind":"listed","path":"scoring.py","file_url":"https://github.com/josephch405/curriculum-nmt/blob/HEAD/scoring.py","link_basis":"first_harvest_node","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"e40cffce14ff31fa"}},{"code_sha256_prefix":"8bc22cc9a2084869","entry":"linear","repo":"josephch405/curriculum-nmt","repo_kind":"listed","path":"pacing.py","file_url":"https://github.com/josephch405/curriculum-nmt/blob/HEAD/pacing.py","link_basis":"first_harvest_node","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"8bc22cc9a2084869"}},{"code_sha256_prefix":"58dc26659bb16162","entry":"pacing_data","repo":"josephch405/curriculum-nmt","repo_kind":"listed","path":"pacing.py","file_url":"https://github.com/josephch405/curriculum-nmt/blob/HEAD/pacing.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"58dc26659bb16162"}},{"code_sha256_prefix":"9d039c026a2b3889","entry":"rank_scores","repo":"josephch405/curriculum-nmt","repo_kind":"listed","path":"scoring.py","file_url":"https://github.com/josephch405/curriculum-nmt/blob/HEAD/scoring.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"9d039c026a2b3889"}},{"code_sha256_prefix":"6c0f0a33d3ea878f","entry":"rarity_scores","repo":"josephch405/curriculum-nmt","repo_kind":"listed","path":"scoring.py","file_url":"https://github.com/josephch405/curriculum-nmt/blob/HEAD/scoring.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"6c0f0a33d3ea878f"}},{"code_sha256_prefix":"3d9f5deeb1f61102","entry":"root","repo":"josephch405/curriculum-nmt","repo_kind":"listed","path":"pacing.py","file_url":"https://github.com/josephch405/curriculum-nmt/blob/HEAD/pacing.py","link_basis":"first_harvest_node","language":"python","status":"ran_honours","verification_level":1,"contract_check":"HONOURS","metamorphic_tier":"well_formed","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"3d9f5deeb1f61102"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}