{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:E2XFO5V2WMXYW4SPD7D4MWIJAH","short_pith_number":"pith:E2XFO5V2","schema_version":"1.0","canonical_sha256":"26ae5776bab32f8b724f1fc7c6590901e9991466f761ad3755e1982184fbf935","source":{"kind":"arxiv","id":"2005.00965","version":1},"attestation_state":"computed","paper":{"title":"Double-Hard Debias: Tailoring Word Embeddings for Gender Bias Mitigation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Bryan McCann, Caiming Xiong, Nazneen Fatema Rajani, Tianlu Wang, Vicente Ordonez, Xi Victoria Lin","submitted_at":"2020-05-03T02:33:20Z","abstract_excerpt":"Word embeddings derived from human-generated corpora inherit strong gender bias which can be further amplified by downstream models. Some commonly adopted debiasing approaches, including the seminal Hard Debias algorithm, apply post-processing procedures that project pre-trained word embeddings into a subspace orthogonal to an inferred gender subspace. We discover that semantic-agnostic corpus regularities such as word frequency captured by the word embeddings negatively impact the performance of these algorithms. We propose a simple but effective technique, Double Hard Debias, which purifies "},"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":"2005.00965","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2020-05-03T02:33:20Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"6d55c34ddea290b94c8eff640ae957373f1437032caf48a76178eb0053d459e5","abstract_canon_sha256":"b71667f1169a8ae9e8ff231609410acc2a390f54210c4d119c3599b5b83fde03"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:00:02.704404Z","signature_b64":"1qwDYmWUYlqpM6ZTKPTrH4r3FHPFeDeTrfQahYk34K13mWYMr+wfxPsmPOkFarADArkYS20THbTtPvMZqJP8CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"26ae5776bab32f8b724f1fc7c6590901e9991466f761ad3755e1982184fbf935","last_reissued_at":"2026-07-05T01:00:02.703926Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:00:02.703926Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Double-Hard Debias: Tailoring Word Embeddings for Gender Bias Mitigation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Bryan McCann, Caiming Xiong, Nazneen Fatema Rajani, Tianlu Wang, Vicente Ordonez, Xi Victoria Lin","submitted_at":"2020-05-03T02:33:20Z","abstract_excerpt":"Word embeddings derived from human-generated corpora inherit strong gender bias which can be further amplified by downstream models. Some commonly adopted debiasing approaches, including the seminal Hard Debias algorithm, apply post-processing procedures that project pre-trained word embeddings into a subspace orthogonal to an inferred gender subspace. We discover that semantic-agnostic corpus regularities such as word frequency captured by the word embeddings negatively impact the performance of these algorithms. We propose a simple but effective technique, Double Hard Debias, which purifies "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2005.00965","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/2005.00965/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":"2005.00965","created_at":"2026-07-05T01:00:02.703985+00:00"},{"alias_kind":"arxiv_version","alias_value":"2005.00965v1","created_at":"2026-07-05T01:00:02.703985+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2005.00965","created_at":"2026-07-05T01:00:02.703985+00:00"},{"alias_kind":"pith_short_12","alias_value":"E2XFO5V2WMXY","created_at":"2026-07-05T01:00:02.703985+00:00"},{"alias_kind":"pith_short_16","alias_value":"E2XFO5V2WMXYW4SP","created_at":"2026-07-05T01:00:02.703985+00:00"},{"alias_kind":"pith_short_8","alias_value":"E2XFO5V2","created_at":"2026-07-05T01:00:02.703985+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2411.12074","citing_title":"Mitigating Gender Bias in Contextual Word Embeddings","ref_index":22,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/E2XFO5V2WMXYW4SPD7D4MWIJAH","json":"https://pith.science/pith/E2XFO5V2WMXYW4SPD7D4MWIJAH.json","graph_json":"https://pith.science/api/pith-number/E2XFO5V2WMXYW4SPD7D4MWIJAH/graph.json","events_json":"https://pith.science/api/pith-number/E2XFO5V2WMXYW4SPD7D4MWIJAH/events.json","paper":"https://pith.science/paper/E2XFO5V2"},"agent_actions":{"view_html":"https://pith.science/pith/E2XFO5V2WMXYW4SPD7D4MWIJAH","download_json":"https://pith.science/pith/E2XFO5V2WMXYW4SPD7D4MWIJAH.json","view_paper":"https://pith.science/paper/E2XFO5V2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2005.00965&json=true","fetch_graph":"https://pith.science/api/pith-number/E2XFO5V2WMXYW4SPD7D4MWIJAH/graph.json","fetch_events":"https://pith.science/api/pith-number/E2XFO5V2WMXYW4SPD7D4MWIJAH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/E2XFO5V2WMXYW4SPD7D4MWIJAH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/E2XFO5V2WMXYW4SPD7D4MWIJAH/action/storage_attestation","attest_author":"https://pith.science/pith/E2XFO5V2WMXYW4SPD7D4MWIJAH/action/author_attestation","sign_citation":"https://pith.science/pith/E2XFO5V2WMXYW4SPD7D4MWIJAH/action/citation_signature","submit_replication":"https://pith.science/pith/E2XFO5V2WMXYW4SPD7D4MWIJAH/action/replication_record"}},"created_at":"2026-07-05T01:00:02.703985+00:00","updated_at":"2026-07-05T01:00:02.703985+00:00"}