{"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/synthetic-and-natural-noise-both-break-neural","title":"Synthetic and Natural Noise Both Break Neural Machine Translation","arxiv_id":"1711.02173","date":"2017-11-06","proceeding":"ICLR 2018 1","authors":["Yonatan Belinkov","Yonatan Bisk"],"abstract":"Character-based neural machine translation (NMT) models alleviate\nout-of-vocabulary issues, learn morphology, and move us closer to completely\nend-to-end translation systems. Unfortunately, they are also very brittle and\neasily falter when presented with noisy data. In this paper, we confront NMT\nmodels with synthetic and natural sources of noise. We find that\nstate-of-the-art models fail to translate even moderately noisy texts that\nhumans have no trouble comprehending. We explore two approaches to increase\nmodel robustness: structure-invariant word representations and robust training\non noisy texts. We find that a model based on a character convolutional neural\nnetwork is able to simultaneously learn representations robust to multiple\nkinds of noise.","url_abs":"http://arxiv.org/abs/1711.02173v2","url_pdf":"http://arxiv.org/pdf/1711.02173v2.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":"synthetic-and-natural-noise-both-break-neural","repo_url":"https://github.com/ybisk/charNMT-noise","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"synthetic-and-natural-noise-both-break-neural","repo_url":"https://github.com/textflint/textflint","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"synthetic-and-natural-noise-both-break-neural","repo_url":"https://github.com/makcedward/nlpaug","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"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":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1711.02173","atlas_url":"https://app.syntology.ai/?focus=1711.02173","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1711.02173"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+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/textflint/textflint","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/makcedward/nlpaug","reach":null},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/ybisk/charNMT-noise","reach":null}],"summary":{"ran_draft_wrong":3},"by_repo_kind":{"official":{"samples":3,"ran":3,"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":3,"samples":[{"code_sha256_prefix":"d7f9d7f194c89efd","entry":"fully_random","repo":"ybisk/charNMT-noise","repo_kind":"official","path":"scrambler.py","file_url":"https://github.com/ybisk/charNMT-noise/blob/HEAD/scrambler.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":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"d7f9d7f194c89efd"}},{"code_sha256_prefix":"33d8867e7d612c53","entry":"random_middle","repo":"ybisk/charNMT-noise","repo_kind":"official","path":"scrambler.py","file_url":"https://github.com/ybisk/charNMT-noise/blob/HEAD/scrambler.py","link_basis":"first_harvest_node","language":"python","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":true,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"33d8867e7d612c53"}},{"code_sha256_prefix":"2c684c93893c4067","entry":"swap","repo":"ybisk/charNMT-noise","repo_kind":"official","path":"scrambler.py","file_url":"https://github.com/ybisk/charNMT-noise/blob/HEAD/scrambler.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":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"2c684c93893c4067"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}