{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:YMPOEIWFMXD46GKQCX5RBDM5DY","short_pith_number":"pith:YMPOEIWF","schema_version":"1.0","canonical_sha256":"c31ee222c565c7cf195015fb108d9d1e04fc8d59a8cb5ac8333e22a8b683fe4b","source":{"kind":"arxiv","id":"2607.03397","version":1},"attestation_state":"computed","paper":{"title":"Efficient bias mitigation in T2I diffusion models using Concept Graphs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.AI","authors_text":"Avinash Kori, Francesco Leofante, Mansi","submitted_at":"2026-07-03T14:58:56Z","abstract_excerpt":"Text-to-Image diffusion models often propagate harmful bias inherited from the training data. Existing bias mitigation techniques typically intervene only at the text encoder or provide inference-time guidance, often leading to generations that collapse into semantically incoherent outputs. To address these limitations, we introduce CO-ALIGN (Concept Ontology Alignment), a novel bias mitigation approach based on concept-graph alignment that operates on the model's internal concept ontology. By aligning concepts within the text encoder and denoiser, CO-ALIGN achieves substantial bias reduction "},"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":"2607.03397","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-07-03T14:58:56Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"dbdc25f50b08c99828c7c0fdec87dd24bc56f9be8059b013967ff663d55ce7a8","abstract_canon_sha256":"021e3a965978163090afa8b164748bc4a4b673cbe2abf8e979b758c5eb7a4d04"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-07T02:17:41.325699Z","signature_b64":"ttqqmOUVRojSsLt5wo3TFIX0viZQKz0cUWVgynqfxSVapQEuwAKzh9p/K78s5EKE8/4bHwzI1NVi6tukh2EaBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c31ee222c565c7cf195015fb108d9d1e04fc8d59a8cb5ac8333e22a8b683fe4b","last_reissued_at":"2026-07-07T02:17:41.324894Z","signature_status":"signed_v1","first_computed_at":"2026-07-07T02:17:41.324894Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Efficient bias mitigation in T2I diffusion models using Concept Graphs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.AI","authors_text":"Avinash Kori, Francesco Leofante, Mansi","submitted_at":"2026-07-03T14:58:56Z","abstract_excerpt":"Text-to-Image diffusion models often propagate harmful bias inherited from the training data. Existing bias mitigation techniques typically intervene only at the text encoder or provide inference-time guidance, often leading to generations that collapse into semantically incoherent outputs. To address these limitations, we introduce CO-ALIGN (Concept Ontology Alignment), a novel bias mitigation approach based on concept-graph alignment that operates on the model's internal concept ontology. By aligning concepts within the text encoder and denoiser, CO-ALIGN achieves substantial bias reduction "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.03397","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/2607.03397/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":"2607.03397","created_at":"2026-07-07T02:17:41.325008+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.03397v1","created_at":"2026-07-07T02:17:41.325008+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.03397","created_at":"2026-07-07T02:17:41.325008+00:00"},{"alias_kind":"pith_short_12","alias_value":"YMPOEIWFMXD4","created_at":"2026-07-07T02:17:41.325008+00:00"},{"alias_kind":"pith_short_16","alias_value":"YMPOEIWFMXD46GKQ","created_at":"2026-07-07T02:17:41.325008+00:00"},{"alias_kind":"pith_short_8","alias_value":"YMPOEIWF","created_at":"2026-07-07T02:17:41.325008+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/YMPOEIWFMXD46GKQCX5RBDM5DY","json":"https://pith.science/pith/YMPOEIWFMXD46GKQCX5RBDM5DY.json","graph_json":"https://pith.science/api/pith-number/YMPOEIWFMXD46GKQCX5RBDM5DY/graph.json","events_json":"https://pith.science/api/pith-number/YMPOEIWFMXD46GKQCX5RBDM5DY/events.json","paper":"https://pith.science/paper/YMPOEIWF"},"agent_actions":{"view_html":"https://pith.science/pith/YMPOEIWFMXD46GKQCX5RBDM5DY","download_json":"https://pith.science/pith/YMPOEIWFMXD46GKQCX5RBDM5DY.json","view_paper":"https://pith.science/paper/YMPOEIWF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.03397&json=true","fetch_graph":"https://pith.science/api/pith-number/YMPOEIWFMXD46GKQCX5RBDM5DY/graph.json","fetch_events":"https://pith.science/api/pith-number/YMPOEIWFMXD46GKQCX5RBDM5DY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YMPOEIWFMXD46GKQCX5RBDM5DY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YMPOEIWFMXD46GKQCX5RBDM5DY/action/storage_attestation","attest_author":"https://pith.science/pith/YMPOEIWFMXD46GKQCX5RBDM5DY/action/author_attestation","sign_citation":"https://pith.science/pith/YMPOEIWFMXD46GKQCX5RBDM5DY/action/citation_signature","submit_replication":"https://pith.science/pith/YMPOEIWFMXD46GKQCX5RBDM5DY/action/replication_record"}},"created_at":"2026-07-07T02:17:41.325008+00:00","updated_at":"2026-07-07T02:17:41.325008+00:00"}