{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:56BENPYJCXLPPILSYT5GHUL4OZ","short_pith_number":"pith:56BENPYJ","schema_version":"1.0","canonical_sha256":"ef8246bf0915d6f7a172c4fa63d17c7676d51c54c1bb82d73b0dc1eeb5cb19ec","source":{"kind":"arxiv","id":"2505.24088","version":1},"attestation_state":"computed","paper":{"title":"Proxy-FDA: Proxy-based Feature Distribution Alignment for Fine-tuning Vision Foundation Models without Forgetting","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Barry-John Theobald, Chen Huang, Fartash Faghri, Hadi Pouransari, Josh Susskind, Mehrdad Farajtabar, Oncel Tuzel, Raviteja Vemulapalli, Skyler Seto","submitted_at":"2025-05-30T00:16:53Z","abstract_excerpt":"Vision foundation models pre-trained on massive data encode rich representations of real-world concepts, which can be adapted to downstream tasks by fine-tuning. However, fine-tuning foundation models on one task often leads to the issue of concept forgetting on other tasks. Recent methods of robust fine-tuning aim to mitigate forgetting of prior knowledge without affecting the fine-tuning performance. Knowledge is often preserved by matching the original and fine-tuned model weights or feature pairs. However, such point-wise matching can be too strong, without explicit awareness of the featur"},"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":"2505.24088","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-05-30T00:16:53Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"d71cdf67e12e8f8d6b8403950e5de5cfdb5abddd52f1bed79e19cfb2d4a0bda4","abstract_canon_sha256":"e1e8b6a7541bf2f26bdfd901c62e7fe973f666331b162033a81c63ac94139714"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:12:41.573765Z","signature_b64":"7etMAc4SHb1Y66ft3metRM5Y96zi/IzqUZ1Z4zbE+H1PexeOAVJb9rqrmKR8f81HNjT/EnchXqJfWbwGsi9yDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ef8246bf0915d6f7a172c4fa63d17c7676d51c54c1bb82d73b0dc1eeb5cb19ec","last_reissued_at":"2026-07-05T11:12:41.573075Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:12:41.573075Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Proxy-FDA: Proxy-based Feature Distribution Alignment for Fine-tuning Vision Foundation Models without Forgetting","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Barry-John Theobald, Chen Huang, Fartash Faghri, Hadi Pouransari, Josh Susskind, Mehrdad Farajtabar, Oncel Tuzel, Raviteja Vemulapalli, Skyler Seto","submitted_at":"2025-05-30T00:16:53Z","abstract_excerpt":"Vision foundation models pre-trained on massive data encode rich representations of real-world concepts, which can be adapted to downstream tasks by fine-tuning. However, fine-tuning foundation models on one task often leads to the issue of concept forgetting on other tasks. Recent methods of robust fine-tuning aim to mitigate forgetting of prior knowledge without affecting the fine-tuning performance. Knowledge is often preserved by matching the original and fine-tuned model weights or feature pairs. However, such point-wise matching can be too strong, without explicit awareness of the featur"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.24088","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/2505.24088/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":"2505.24088","created_at":"2026-07-05T11:12:41.573165+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.24088v1","created_at":"2026-07-05T11:12:41.573165+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.24088","created_at":"2026-07-05T11:12:41.573165+00:00"},{"alias_kind":"pith_short_12","alias_value":"56BENPYJCXLP","created_at":"2026-07-05T11:12:41.573165+00:00"},{"alias_kind":"pith_short_16","alias_value":"56BENPYJCXLPPILS","created_at":"2026-07-05T11:12:41.573165+00:00"},{"alias_kind":"pith_short_8","alias_value":"56BENPYJ","created_at":"2026-07-05T11:12:41.573165+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/56BENPYJCXLPPILSYT5GHUL4OZ","json":"https://pith.science/pith/56BENPYJCXLPPILSYT5GHUL4OZ.json","graph_json":"https://pith.science/api/pith-number/56BENPYJCXLPPILSYT5GHUL4OZ/graph.json","events_json":"https://pith.science/api/pith-number/56BENPYJCXLPPILSYT5GHUL4OZ/events.json","paper":"https://pith.science/paper/56BENPYJ"},"agent_actions":{"view_html":"https://pith.science/pith/56BENPYJCXLPPILSYT5GHUL4OZ","download_json":"https://pith.science/pith/56BENPYJCXLPPILSYT5GHUL4OZ.json","view_paper":"https://pith.science/paper/56BENPYJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.24088&json=true","fetch_graph":"https://pith.science/api/pith-number/56BENPYJCXLPPILSYT5GHUL4OZ/graph.json","fetch_events":"https://pith.science/api/pith-number/56BENPYJCXLPPILSYT5GHUL4OZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/56BENPYJCXLPPILSYT5GHUL4OZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/56BENPYJCXLPPILSYT5GHUL4OZ/action/storage_attestation","attest_author":"https://pith.science/pith/56BENPYJCXLPPILSYT5GHUL4OZ/action/author_attestation","sign_citation":"https://pith.science/pith/56BENPYJCXLPPILSYT5GHUL4OZ/action/citation_signature","submit_replication":"https://pith.science/pith/56BENPYJCXLPPILSYT5GHUL4OZ/action/replication_record"}},"created_at":"2026-07-05T11:12:41.573165+00:00","updated_at":"2026-07-05T11:12:41.573165+00:00"}