{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:XSYBCHI5WQKMDWD2FWFAB2YLSX","short_pith_number":"pith:XSYBCHI5","schema_version":"1.0","canonical_sha256":"bcb0111d1db414c1d87a2d8a00eb0b95c9ced1e3b93814ad0caa6312fc84deea","source":{"kind":"arxiv","id":"2012.04842","version":2},"attestation_state":"computed","paper":{"title":"Improving the Fairness of Deep Generative Models without Retraining","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Bolei Zhou, Shuhan Tan, Yujun Shen","submitted_at":"2020-12-09T03:20:41Z","abstract_excerpt":"Generative Adversarial Networks (GANs) advance face synthesis through learning the underlying distribution of observed data. Despite the high-quality generated faces, some minority groups can be rarely generated from the trained models due to a biased image generation process. To study the issue, we first conduct an empirical study on a pre-trained face synthesis model. We observe that after training the GAN model not only carries the biases in the training data but also amplifies them to some degree in the image generation process. To further improve the fairness of image generation, we propo"},"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":"2012.04842","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2020-12-09T03:20:41Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"bc850dc1e9802b9e32a072cebf6f804694ab227c7dc8dcbd42341469613bb3a7","abstract_canon_sha256":"fbdb2b0dcaa5b1a511816fb9e9b47d3fe41a7e31f8f49965619cbffa92da7f37"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:26:53.101894Z","signature_b64":"7tPHm+QzmtVT3lbSXfXp7sNK/UkAbkiAMC1YyIaLb/xWeVnvlnurirMOmQmpAc6lHhjr0YcnAN7EOJrEOA7mAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bcb0111d1db414c1d87a2d8a00eb0b95c9ced1e3b93814ad0caa6312fc84deea","last_reissued_at":"2026-07-05T02:26:53.101477Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:26:53.101477Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Improving the Fairness of Deep Generative Models without Retraining","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Bolei Zhou, Shuhan Tan, Yujun Shen","submitted_at":"2020-12-09T03:20:41Z","abstract_excerpt":"Generative Adversarial Networks (GANs) advance face synthesis through learning the underlying distribution of observed data. Despite the high-quality generated faces, some minority groups can be rarely generated from the trained models due to a biased image generation process. To study the issue, we first conduct an empirical study on a pre-trained face synthesis model. We observe that after training the GAN model not only carries the biases in the training data but also amplifies them to some degree in the image generation process. To further improve the fairness of image generation, we propo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2012.04842","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/2012.04842/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":"2012.04842","created_at":"2026-07-05T02:26:53.101533+00:00"},{"alias_kind":"arxiv_version","alias_value":"2012.04842v2","created_at":"2026-07-05T02:26:53.101533+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2012.04842","created_at":"2026-07-05T02:26:53.101533+00:00"},{"alias_kind":"pith_short_12","alias_value":"XSYBCHI5WQKM","created_at":"2026-07-05T02:26:53.101533+00:00"},{"alias_kind":"pith_short_16","alias_value":"XSYBCHI5WQKMDWD2","created_at":"2026-07-05T02:26:53.101533+00:00"},{"alias_kind":"pith_short_8","alias_value":"XSYBCHI5","created_at":"2026-07-05T02:26:53.101533+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.09052","citing_title":"Contrastive Conditional-Unconditional Alignment for Long-tailed Diffusion Model","ref_index":40,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XSYBCHI5WQKMDWD2FWFAB2YLSX","json":"https://pith.science/pith/XSYBCHI5WQKMDWD2FWFAB2YLSX.json","graph_json":"https://pith.science/api/pith-number/XSYBCHI5WQKMDWD2FWFAB2YLSX/graph.json","events_json":"https://pith.science/api/pith-number/XSYBCHI5WQKMDWD2FWFAB2YLSX/events.json","paper":"https://pith.science/paper/XSYBCHI5"},"agent_actions":{"view_html":"https://pith.science/pith/XSYBCHI5WQKMDWD2FWFAB2YLSX","download_json":"https://pith.science/pith/XSYBCHI5WQKMDWD2FWFAB2YLSX.json","view_paper":"https://pith.science/paper/XSYBCHI5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2012.04842&json=true","fetch_graph":"https://pith.science/api/pith-number/XSYBCHI5WQKMDWD2FWFAB2YLSX/graph.json","fetch_events":"https://pith.science/api/pith-number/XSYBCHI5WQKMDWD2FWFAB2YLSX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XSYBCHI5WQKMDWD2FWFAB2YLSX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XSYBCHI5WQKMDWD2FWFAB2YLSX/action/storage_attestation","attest_author":"https://pith.science/pith/XSYBCHI5WQKMDWD2FWFAB2YLSX/action/author_attestation","sign_citation":"https://pith.science/pith/XSYBCHI5WQKMDWD2FWFAB2YLSX/action/citation_signature","submit_replication":"https://pith.science/pith/XSYBCHI5WQKMDWD2FWFAB2YLSX/action/replication_record"}},"created_at":"2026-07-05T02:26:53.101533+00:00","updated_at":"2026-07-05T02:26:53.101533+00:00"}