{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:UXPOTZVUXMWA5NZPXG37VNPBRP","short_pith_number":"pith:UXPOTZVU","schema_version":"1.0","canonical_sha256":"a5dee9e6b4bb2c0eb72fb9b7fab5e18bf5caefa995ec9ad2973f42540881316c","source":{"kind":"arxiv","id":"2104.09810","version":1},"attestation_state":"computed","paper":{"title":"Addressing the Vulnerability of NMT in Input Perturbations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Ai Ti Aw, Kui Wu, Shafiq Joty, Weiwen Xu, Yang Ding","submitted_at":"2021-04-20T07:52:58Z","abstract_excerpt":"Neural Machine Translation (NMT) has achieved significant breakthrough in performance but is known to suffer vulnerability to input perturbations. As real input noise is difficult to predict during training, robustness is a big issue for system deployment. In this paper, we improve the robustness of NMT models by reducing the effect of noisy words through a Context-Enhanced Reconstruction (CER) approach. CER trains the model to resist noise in two steps: (1) perturbation step that breaks the naturalness of input sequence with made-up words; (2) reconstruction step that defends the noise propag"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2104.09810","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2021-04-20T07:52:58Z","cross_cats_sorted":[],"title_canon_sha256":"a1a99f99c09f385e78724df2355400c9672c5ea3d39d648113d11b98179be96e","abstract_canon_sha256":"717bd906da3ea96d1f5cfd43d509adc3f49415e9b2743c1e54985daf19ddfce8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:33:37.183521Z","signature_b64":"6gwP+wRzaw9RZQlKt7rPMBDGCG8pdx4E0cXCXhfbsd7vjUMy6i0NAEIVEwmGaUVsrvHKsXl9eNCMaJE2pYftDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a5dee9e6b4bb2c0eb72fb9b7fab5e18bf5caefa995ec9ad2973f42540881316c","last_reissued_at":"2026-07-05T02:33:37.183099Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:33:37.183099Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Addressing the Vulnerability of NMT in Input Perturbations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Ai Ti Aw, Kui Wu, Shafiq Joty, Weiwen Xu, Yang Ding","submitted_at":"2021-04-20T07:52:58Z","abstract_excerpt":"Neural Machine Translation (NMT) has achieved significant breakthrough in performance but is known to suffer vulnerability to input perturbations. As real input noise is difficult to predict during training, robustness is a big issue for system deployment. In this paper, we improve the robustness of NMT models by reducing the effect of noisy words through a Context-Enhanced Reconstruction (CER) approach. CER trains the model to resist noise in two steps: (1) perturbation step that breaks the naturalness of input sequence with made-up words; (2) reconstruction step that defends the noise propag"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2104.09810","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2104.09810/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2104.09810","created_at":"2026-07-05T02:33:37.183161+00:00"},{"alias_kind":"arxiv_version","alias_value":"2104.09810v1","created_at":"2026-07-05T02:33:37.183161+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2104.09810","created_at":"2026-07-05T02:33:37.183161+00:00"},{"alias_kind":"pith_short_12","alias_value":"UXPOTZVUXMWA","created_at":"2026-07-05T02:33:37.183161+00:00"},{"alias_kind":"pith_short_16","alias_value":"UXPOTZVUXMWA5NZP","created_at":"2026-07-05T02:33:37.183161+00:00"},{"alias_kind":"pith_short_8","alias_value":"UXPOTZVU","created_at":"2026-07-05T02:33:37.183161+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UXPOTZVUXMWA5NZPXG37VNPBRP","json":"https://pith.science/pith/UXPOTZVUXMWA5NZPXG37VNPBRP.json","graph_json":"https://pith.science/api/pith-number/UXPOTZVUXMWA5NZPXG37VNPBRP/graph.json","events_json":"https://pith.science/api/pith-number/UXPOTZVUXMWA5NZPXG37VNPBRP/events.json","paper":"https://pith.science/paper/UXPOTZVU"},"agent_actions":{"view_html":"https://pith.science/pith/UXPOTZVUXMWA5NZPXG37VNPBRP","download_json":"https://pith.science/pith/UXPOTZVUXMWA5NZPXG37VNPBRP.json","view_paper":"https://pith.science/paper/UXPOTZVU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2104.09810&json=true","fetch_graph":"https://pith.science/api/pith-number/UXPOTZVUXMWA5NZPXG37VNPBRP/graph.json","fetch_events":"https://pith.science/api/pith-number/UXPOTZVUXMWA5NZPXG37VNPBRP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UXPOTZVUXMWA5NZPXG37VNPBRP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UXPOTZVUXMWA5NZPXG37VNPBRP/action/storage_attestation","attest_author":"https://pith.science/pith/UXPOTZVUXMWA5NZPXG37VNPBRP/action/author_attestation","sign_citation":"https://pith.science/pith/UXPOTZVUXMWA5NZPXG37VNPBRP/action/citation_signature","submit_replication":"https://pith.science/pith/UXPOTZVUXMWA5NZPXG37VNPBRP/action/replication_record"}},"created_at":"2026-07-05T02:33:37.183161+00:00","updated_at":"2026-07-05T02:33:37.183161+00:00"}