{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:FW37J22CVLTWEUOG6CMEM7HURQ","short_pith_number":"pith:FW37J22C","schema_version":"1.0","canonical_sha256":"2db7f4eb42aae76251c6f098467cf48c32b678905a094bf27d78063c06433c8e","source":{"kind":"arxiv","id":"2101.09523","version":1},"attestation_state":"computed","paper":{"title":"Debiasing Pre-trained Contextualised Embeddings","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Danushka Bollegala, Masahiro Kaneko","submitted_at":"2021-01-23T15:28:48Z","abstract_excerpt":"In comparison to the numerous debiasing methods proposed for the static non-contextualised word embeddings, the discriminative biases in contextualised embeddings have received relatively little attention. We propose a fine-tuning method that can be applied at token- or sentence-levels to debias pre-trained contextualised embeddings. Our proposed method can be applied to any pre-trained contextualised embedding model, without requiring to retrain those models. Using gender bias as an illustrative example, we then conduct a systematic study using several state-of-the-art (SoTA) contextualised r"},"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":"2101.09523","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2021-01-23T15:28:48Z","cross_cats_sorted":[],"title_canon_sha256":"c3e7774292c8c9348bd25bf028237f10587f0b0a9911e9892d9e428e382af04d","abstract_canon_sha256":"5d2959e0345e5db33a33104118a930be2a31d8d944e4bbbf209c5b53bbc65ce1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:09:19.598939Z","signature_b64":"vy2lxLGElr9VUk8hBWSUziOgBwwb+8Olcua2g98WqxkxzlxNSS4TQ4qaAu8bpit2N0FHXJTRJYKAlcbRogkKAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2db7f4eb42aae76251c6f098467cf48c32b678905a094bf27d78063c06433c8e","last_reissued_at":"2026-07-05T02:09:19.598545Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:09:19.598545Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Debiasing Pre-trained Contextualised Embeddings","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Danushka Bollegala, Masahiro Kaneko","submitted_at":"2021-01-23T15:28:48Z","abstract_excerpt":"In comparison to the numerous debiasing methods proposed for the static non-contextualised word embeddings, the discriminative biases in contextualised embeddings have received relatively little attention. We propose a fine-tuning method that can be applied at token- or sentence-levels to debias pre-trained contextualised embeddings. Our proposed method can be applied to any pre-trained contextualised embedding model, without requiring to retrain those models. Using gender bias as an illustrative example, we then conduct a systematic study using several state-of-the-art (SoTA) contextualised r"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2101.09523","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/2101.09523/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":"2101.09523","created_at":"2026-07-05T02:09:19.598610+00:00"},{"alias_kind":"arxiv_version","alias_value":"2101.09523v1","created_at":"2026-07-05T02:09:19.598610+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2101.09523","created_at":"2026-07-05T02:09:19.598610+00:00"},{"alias_kind":"pith_short_12","alias_value":"FW37J22CVLTW","created_at":"2026-07-05T02:09:19.598610+00:00"},{"alias_kind":"pith_short_16","alias_value":"FW37J22CVLTWEUOG","created_at":"2026-07-05T02:09:19.598610+00:00"},{"alias_kind":"pith_short_8","alias_value":"FW37J22C","created_at":"2026-07-05T02:09:19.598610+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.05610","citing_title":"Mitigating Confounding in Speech-Based Dementia Detection through Weight Masking","ref_index":23,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FW37J22CVLTWEUOG6CMEM7HURQ","json":"https://pith.science/pith/FW37J22CVLTWEUOG6CMEM7HURQ.json","graph_json":"https://pith.science/api/pith-number/FW37J22CVLTWEUOG6CMEM7HURQ/graph.json","events_json":"https://pith.science/api/pith-number/FW37J22CVLTWEUOG6CMEM7HURQ/events.json","paper":"https://pith.science/paper/FW37J22C"},"agent_actions":{"view_html":"https://pith.science/pith/FW37J22CVLTWEUOG6CMEM7HURQ","download_json":"https://pith.science/pith/FW37J22CVLTWEUOG6CMEM7HURQ.json","view_paper":"https://pith.science/paper/FW37J22C","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2101.09523&json=true","fetch_graph":"https://pith.science/api/pith-number/FW37J22CVLTWEUOG6CMEM7HURQ/graph.json","fetch_events":"https://pith.science/api/pith-number/FW37J22CVLTWEUOG6CMEM7HURQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FW37J22CVLTWEUOG6CMEM7HURQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FW37J22CVLTWEUOG6CMEM7HURQ/action/storage_attestation","attest_author":"https://pith.science/pith/FW37J22CVLTWEUOG6CMEM7HURQ/action/author_attestation","sign_citation":"https://pith.science/pith/FW37J22CVLTWEUOG6CMEM7HURQ/action/citation_signature","submit_replication":"https://pith.science/pith/FW37J22CVLTWEUOG6CMEM7HURQ/action/replication_record"}},"created_at":"2026-07-05T02:09:19.598610+00:00","updated_at":"2026-07-05T02:09:19.598610+00:00"}