{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:3OBG4A6MQ3BSW5URZ2Y2R3CST3","short_pith_number":"pith:3OBG4A6M","schema_version":"1.0","canonical_sha256":"db826e03cc86c32b7691ceb1a8ec529edd93abeed2a9f00684b7080bf23ea017","source":{"kind":"arxiv","id":"2404.11536","version":2},"attestation_state":"computed","paper":{"title":"FedPFT: Federated Proxy Fine-Tuning of Foundation Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Chenglu Wen, Cheng Wang, Ruisheng Zhang, Shirui Pan, Xiaoliang Fan, Yufan Chen, Zhaopeng Peng, Zheng Wang","submitted_at":"2024-04-17T16:30:06Z","abstract_excerpt":"Adapting Foundation Models (FMs) for downstream tasks through Federated Learning (FL) emerges a promising strategy for protecting data privacy and valuable FMs. Existing methods fine-tune FM by allocating sub-FM to clients in FL, however, leading to suboptimal performance due to insufficient tuning and inevitable error accumulations of gradients. In this paper, we propose Federated Proxy Fine-Tuning (FedPFT), a novel method enhancing FMs adaptation in downstream tasks through FL by two key modules. First, the sub-FM construction module employs a layer-wise compression approach, facilitating co"},"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":"2404.11536","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-04-17T16:30:06Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"3353dcbc168831ef546d40649a1f0c46c6630e60c1b665bd9901dcce07ae10f7","abstract_canon_sha256":"a97174518494a884267ea03b9aa022855f0fb32ca71f661ffa762dad79c7b5f5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:12:50.192939Z","signature_b64":"yGQX9gRDQnXhXZyQzjOjdfd+AiVkIcZDL1sp3Kyva5CDBneMiiqzGAv4pA367//mbbCiJvt/upshhxZdJH0PCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"db826e03cc86c32b7691ceb1a8ec529edd93abeed2a9f00684b7080bf23ea017","last_reissued_at":"2026-07-05T08:12:50.192473Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:12:50.192473Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FedPFT: Federated Proxy Fine-Tuning of Foundation Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Chenglu Wen, Cheng Wang, Ruisheng Zhang, Shirui Pan, Xiaoliang Fan, Yufan Chen, Zhaopeng Peng, Zheng Wang","submitted_at":"2024-04-17T16:30:06Z","abstract_excerpt":"Adapting Foundation Models (FMs) for downstream tasks through Federated Learning (FL) emerges a promising strategy for protecting data privacy and valuable FMs. Existing methods fine-tune FM by allocating sub-FM to clients in FL, however, leading to suboptimal performance due to insufficient tuning and inevitable error accumulations of gradients. In this paper, we propose Federated Proxy Fine-Tuning (FedPFT), a novel method enhancing FMs adaptation in downstream tasks through FL by two key modules. First, the sub-FM construction module employs a layer-wise compression approach, facilitating co"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.11536","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/2404.11536/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":"2404.11536","created_at":"2026-07-05T08:12:50.192527+00:00"},{"alias_kind":"arxiv_version","alias_value":"2404.11536v2","created_at":"2026-07-05T08:12:50.192527+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.11536","created_at":"2026-07-05T08:12:50.192527+00:00"},{"alias_kind":"pith_short_12","alias_value":"3OBG4A6MQ3BS","created_at":"2026-07-05T08:12:50.192527+00:00"},{"alias_kind":"pith_short_16","alias_value":"3OBG4A6MQ3BSW5UR","created_at":"2026-07-05T08:12:50.192527+00:00"},{"alias_kind":"pith_short_8","alias_value":"3OBG4A6M","created_at":"2026-07-05T08:12:50.192527+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.01194","citing_title":"FedRPCA: Enhancing Federated LoRA Aggregation Using Robust PCA","ref_index":55,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3OBG4A6MQ3BSW5URZ2Y2R3CST3","json":"https://pith.science/pith/3OBG4A6MQ3BSW5URZ2Y2R3CST3.json","graph_json":"https://pith.science/api/pith-number/3OBG4A6MQ3BSW5URZ2Y2R3CST3/graph.json","events_json":"https://pith.science/api/pith-number/3OBG4A6MQ3BSW5URZ2Y2R3CST3/events.json","paper":"https://pith.science/paper/3OBG4A6M"},"agent_actions":{"view_html":"https://pith.science/pith/3OBG4A6MQ3BSW5URZ2Y2R3CST3","download_json":"https://pith.science/pith/3OBG4A6MQ3BSW5URZ2Y2R3CST3.json","view_paper":"https://pith.science/paper/3OBG4A6M","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2404.11536&json=true","fetch_graph":"https://pith.science/api/pith-number/3OBG4A6MQ3BSW5URZ2Y2R3CST3/graph.json","fetch_events":"https://pith.science/api/pith-number/3OBG4A6MQ3BSW5URZ2Y2R3CST3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3OBG4A6MQ3BSW5URZ2Y2R3CST3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3OBG4A6MQ3BSW5URZ2Y2R3CST3/action/storage_attestation","attest_author":"https://pith.science/pith/3OBG4A6MQ3BSW5URZ2Y2R3CST3/action/author_attestation","sign_citation":"https://pith.science/pith/3OBG4A6MQ3BSW5URZ2Y2R3CST3/action/citation_signature","submit_replication":"https://pith.science/pith/3OBG4A6MQ3BSW5URZ2Y2R3CST3/action/replication_record"}},"created_at":"2026-07-05T08:12:50.192527+00:00","updated_at":"2026-07-05T08:12:50.192527+00:00"}