{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:N3TDHIXG7TGHV2YLUA5VEVCWT2","short_pith_number":"pith:N3TDHIXG","schema_version":"1.0","canonical_sha256":"6ee633a2e6fccc7aeb0ba03b5254569eb1af937a4ad1196375693369cb17b859","source":{"kind":"arxiv","id":"1908.10763","version":2},"attestation_state":"computed","paper":{"title":"Unlearn Dataset Bias in Natural Language Inference by Fitting the Residual","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Haohan Wang, He He, Sheng Zha","submitted_at":"2019-08-28T15:02:45Z","abstract_excerpt":"Statistical natural language inference (NLI) models are susceptible to learning dataset bias: superficial cues that happen to associate with the label on a particular dataset, but are not useful in general, e.g., negation words indicate contradiction. As exposed by several recent challenge datasets, these models perform poorly when such association is absent, e.g., predicting that \"I love dogs\" contradicts \"I don't love cats\". Our goal is to design learning algorithms that guard against known dataset bias. We formalize the concept of dataset bias under the framework of distribution shift and p"},"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":"1908.10763","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-08-28T15:02:45Z","cross_cats_sorted":[],"title_canon_sha256":"ecc9dd45c095516f7b64292e45dbc6af5ea555b8ebaa37fdafc44d2c9d9a5c4b","abstract_canon_sha256":"aa18629e5731bfcdf8e02f963c05cb0fb39ab55995331cd8a4117bae8087eb66"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:21:36.393523Z","signature_b64":"yudvqEghGPrnsVw1ykQJ6D9jPNaR+xgRZ2VPtJwA0zTTCD2rqeC2E7vOp1ypNa2iaPgL1t3h5BWk5THK+7uABQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6ee633a2e6fccc7aeb0ba03b5254569eb1af937a4ad1196375693369cb17b859","last_reissued_at":"2026-07-05T00:21:36.392951Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:21:36.392951Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Unlearn Dataset Bias in Natural Language Inference by Fitting the Residual","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Haohan Wang, He He, Sheng Zha","submitted_at":"2019-08-28T15:02:45Z","abstract_excerpt":"Statistical natural language inference (NLI) models are susceptible to learning dataset bias: superficial cues that happen to associate with the label on a particular dataset, but are not useful in general, e.g., negation words indicate contradiction. As exposed by several recent challenge datasets, these models perform poorly when such association is absent, e.g., predicting that \"I love dogs\" contradicts \"I don't love cats\". Our goal is to design learning algorithms that guard against known dataset bias. We formalize the concept of dataset bias under the framework of distribution shift and p"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1908.10763","kind":"arxiv","version":2},"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/1908.10763/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":"1908.10763","created_at":"2026-07-05T00:21:36.393020+00:00"},{"alias_kind":"arxiv_version","alias_value":"1908.10763v2","created_at":"2026-07-05T00:21:36.393020+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1908.10763","created_at":"2026-07-05T00:21:36.393020+00:00"},{"alias_kind":"pith_short_12","alias_value":"N3TDHIXG7TGH","created_at":"2026-07-05T00:21:36.393020+00:00"},{"alias_kind":"pith_short_16","alias_value":"N3TDHIXG7TGHV2YL","created_at":"2026-07-05T00:21:36.393020+00:00"},{"alias_kind":"pith_short_8","alias_value":"N3TDHIXG","created_at":"2026-07-05T00:21:36.393020+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/N3TDHIXG7TGHV2YLUA5VEVCWT2","json":"https://pith.science/pith/N3TDHIXG7TGHV2YLUA5VEVCWT2.json","graph_json":"https://pith.science/api/pith-number/N3TDHIXG7TGHV2YLUA5VEVCWT2/graph.json","events_json":"https://pith.science/api/pith-number/N3TDHIXG7TGHV2YLUA5VEVCWT2/events.json","paper":"https://pith.science/paper/N3TDHIXG"},"agent_actions":{"view_html":"https://pith.science/pith/N3TDHIXG7TGHV2YLUA5VEVCWT2","download_json":"https://pith.science/pith/N3TDHIXG7TGHV2YLUA5VEVCWT2.json","view_paper":"https://pith.science/paper/N3TDHIXG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1908.10763&json=true","fetch_graph":"https://pith.science/api/pith-number/N3TDHIXG7TGHV2YLUA5VEVCWT2/graph.json","fetch_events":"https://pith.science/api/pith-number/N3TDHIXG7TGHV2YLUA5VEVCWT2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/N3TDHIXG7TGHV2YLUA5VEVCWT2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/N3TDHIXG7TGHV2YLUA5VEVCWT2/action/storage_attestation","attest_author":"https://pith.science/pith/N3TDHIXG7TGHV2YLUA5VEVCWT2/action/author_attestation","sign_citation":"https://pith.science/pith/N3TDHIXG7TGHV2YLUA5VEVCWT2/action/citation_signature","submit_replication":"https://pith.science/pith/N3TDHIXG7TGHV2YLUA5VEVCWT2/action/replication_record"}},"created_at":"2026-07-05T00:21:36.393020+00:00","updated_at":"2026-07-05T00:21:36.393020+00:00"}