{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:DJGXBWIDJKYVKI3DF7CWK7LICF","short_pith_number":"pith:DJGXBWID","schema_version":"1.0","canonical_sha256":"1a4d70d9034ab15523632fc5657d68117eeb13b00b5ec757f8eefd012cd7b0dc","source":{"kind":"arxiv","id":"2305.01397","version":3},"attestation_state":"computed","paper":{"title":"Are demographically invariant models and representations in medical imaging fair?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CY","eess.IV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Aasa Feragen, Eike Petersen, Enzo Ferrante, Melanie Ganz","submitted_at":"2023-05-02T13:16:04Z","abstract_excerpt":"Medical imaging models have been shown to encode information about patient demographics such as age, race, and sex in their latent representation, raising concerns about their potential for discrimination. Here, we ask whether requiring models not to encode demographic attributes is desirable. We point out that marginal and class-conditional representation invariance imply the standard group fairness notions of demographic parity and equalized odds, respectively. In addition, however, they require matching the risk distributions, thus potentially equalizing away important group differences. En"},"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":"2305.01397","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-05-02T13:16:04Z","cross_cats_sorted":["cs.CY","eess.IV","stat.ML"],"title_canon_sha256":"2fc6f510db96a1b0a2e781925fcba290b59d904a3224d28ef56401750fc1fd91","abstract_canon_sha256":"8321e72f6c01840c0eff77b09acba1a48bf0408ada5ae409e0ec7dc8651e9f64"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:39:26.883222Z","signature_b64":"z/0yflQ+Yjdro652vXvuW2a4VYh/7cVfNQ6uc607zH1EPhIToSqXSqovfTLXOVD9Itz96IQBwHOeGriQelPaBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1a4d70d9034ab15523632fc5657d68117eeb13b00b5ec757f8eefd012cd7b0dc","last_reissued_at":"2026-07-05T08:39:26.882802Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:39:26.882802Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Are demographically invariant models and representations in medical imaging fair?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CY","eess.IV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Aasa Feragen, Eike Petersen, Enzo Ferrante, Melanie Ganz","submitted_at":"2023-05-02T13:16:04Z","abstract_excerpt":"Medical imaging models have been shown to encode information about patient demographics such as age, race, and sex in their latent representation, raising concerns about their potential for discrimination. Here, we ask whether requiring models not to encode demographic attributes is desirable. We point out that marginal and class-conditional representation invariance imply the standard group fairness notions of demographic parity and equalized odds, respectively. In addition, however, they require matching the risk distributions, thus potentially equalizing away important group differences. En"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.01397","kind":"arxiv","version":3},"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/2305.01397/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":"2305.01397","created_at":"2026-07-05T08:39:26.882858+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.01397v3","created_at":"2026-07-05T08:39:26.882858+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.01397","created_at":"2026-07-05T08:39:26.882858+00:00"},{"alias_kind":"pith_short_12","alias_value":"DJGXBWIDJKYV","created_at":"2026-07-05T08:39:26.882858+00:00"},{"alias_kind":"pith_short_16","alias_value":"DJGXBWIDJKYVKI3D","created_at":"2026-07-05T08:39:26.882858+00:00"},{"alias_kind":"pith_short_8","alias_value":"DJGXBWID","created_at":"2026-07-05T08:39:26.882858+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.06891","citing_title":"Towards Fairness under Label Bias in Image Segmentation: Impact, Measurement and Mitigation","ref_index":9,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DJGXBWIDJKYVKI3DF7CWK7LICF","json":"https://pith.science/pith/DJGXBWIDJKYVKI3DF7CWK7LICF.json","graph_json":"https://pith.science/api/pith-number/DJGXBWIDJKYVKI3DF7CWK7LICF/graph.json","events_json":"https://pith.science/api/pith-number/DJGXBWIDJKYVKI3DF7CWK7LICF/events.json","paper":"https://pith.science/paper/DJGXBWID"},"agent_actions":{"view_html":"https://pith.science/pith/DJGXBWIDJKYVKI3DF7CWK7LICF","download_json":"https://pith.science/pith/DJGXBWIDJKYVKI3DF7CWK7LICF.json","view_paper":"https://pith.science/paper/DJGXBWID","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.01397&json=true","fetch_graph":"https://pith.science/api/pith-number/DJGXBWIDJKYVKI3DF7CWK7LICF/graph.json","fetch_events":"https://pith.science/api/pith-number/DJGXBWIDJKYVKI3DF7CWK7LICF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DJGXBWIDJKYVKI3DF7CWK7LICF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DJGXBWIDJKYVKI3DF7CWK7LICF/action/storage_attestation","attest_author":"https://pith.science/pith/DJGXBWIDJKYVKI3DF7CWK7LICF/action/author_attestation","sign_citation":"https://pith.science/pith/DJGXBWIDJKYVKI3DF7CWK7LICF/action/citation_signature","submit_replication":"https://pith.science/pith/DJGXBWIDJKYVKI3DF7CWK7LICF/action/replication_record"}},"created_at":"2026-07-05T08:39:26.882858+00:00","updated_at":"2026-07-05T08:39:26.882858+00:00"}