{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:NOKRD3ZLF7WJFJ7XS2SO5HRQRJ","short_pith_number":"pith:NOKRD3ZL","schema_version":"1.0","canonical_sha256":"6b9511ef2b2fec92a7f796a4ee9e308a48063fd9f05099d755422e5a61e00c53","source":{"kind":"arxiv","id":"2502.06376","version":3},"attestation_state":"computed","paper":{"title":"Many-Task Federated Fine-Tuning via Unified Task Vectors","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Nirvana Meratnia, Tanir Ozcelebi, Vasileios Tsouvalas","submitted_at":"2025-02-10T11:56:02Z","abstract_excerpt":"Federated Learning (FL) traditionally assumes homogeneous client tasks; however, in real-world scenarios, clients often specialize in diverse tasks, introducing task heterogeneity. To address this challenge, Many-Task FL (MaT-FL) has emerged, enabling clients to collaborate effectively despite task diversity. Existing MaT-FL approaches rely on client grouping or personalized layers, requiring the server to manage individual models and failing to account for clients handling multiple tasks. We propose MaTU, a MaT-FL approach that enables joint learning of task vectors across clients, eliminatin"},"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":"2502.06376","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-10T11:56:02Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"a138da385edd8509e6ddb75d8401b10fbc3ea42083153dbff87e02030d0370c8","abstract_canon_sha256":"a28cf9fc5da53d911db9d5f180c888da6b8e55061368f1ce1d89ea651f33b63c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:34:04.347495Z","signature_b64":"kBD2hTX7q+3Re7c+sM+eMbu4DlVSGqwSd1nyPmESagCcWZYNS7ytV4BlZUTiHosA3KgUHQztHUaNzDPCuH4WCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6b9511ef2b2fec92a7f796a4ee9e308a48063fd9f05099d755422e5a61e00c53","last_reissued_at":"2026-07-05T11:34:04.347067Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:34:04.347067Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Many-Task Federated Fine-Tuning via Unified Task Vectors","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Nirvana Meratnia, Tanir Ozcelebi, Vasileios Tsouvalas","submitted_at":"2025-02-10T11:56:02Z","abstract_excerpt":"Federated Learning (FL) traditionally assumes homogeneous client tasks; however, in real-world scenarios, clients often specialize in diverse tasks, introducing task heterogeneity. To address this challenge, Many-Task FL (MaT-FL) has emerged, enabling clients to collaborate effectively despite task diversity. Existing MaT-FL approaches rely on client grouping or personalized layers, requiring the server to manage individual models and failing to account for clients handling multiple tasks. We propose MaTU, a MaT-FL approach that enables joint learning of task vectors across clients, eliminatin"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.06376","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/2502.06376/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":"2502.06376","created_at":"2026-07-05T11:34:04.347123+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.06376v3","created_at":"2026-07-05T11:34:04.347123+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.06376","created_at":"2026-07-05T11:34:04.347123+00:00"},{"alias_kind":"pith_short_12","alias_value":"NOKRD3ZLF7WJ","created_at":"2026-07-05T11:34:04.347123+00:00"},{"alias_kind":"pith_short_16","alias_value":"NOKRD3ZLF7WJFJ7X","created_at":"2026-07-05T11:34:04.347123+00:00"},{"alias_kind":"pith_short_8","alias_value":"NOKRD3ZL","created_at":"2026-07-05T11:34:04.347123+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2501.09934","citing_title":"HEART: Achieving Timely Multi-Model Training for Vehicle-Edge-Cloud-Integrated Hierarchical Federated Learning","ref_index":21,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NOKRD3ZLF7WJFJ7XS2SO5HRQRJ","json":"https://pith.science/pith/NOKRD3ZLF7WJFJ7XS2SO5HRQRJ.json","graph_json":"https://pith.science/api/pith-number/NOKRD3ZLF7WJFJ7XS2SO5HRQRJ/graph.json","events_json":"https://pith.science/api/pith-number/NOKRD3ZLF7WJFJ7XS2SO5HRQRJ/events.json","paper":"https://pith.science/paper/NOKRD3ZL"},"agent_actions":{"view_html":"https://pith.science/pith/NOKRD3ZLF7WJFJ7XS2SO5HRQRJ","download_json":"https://pith.science/pith/NOKRD3ZLF7WJFJ7XS2SO5HRQRJ.json","view_paper":"https://pith.science/paper/NOKRD3ZL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.06376&json=true","fetch_graph":"https://pith.science/api/pith-number/NOKRD3ZLF7WJFJ7XS2SO5HRQRJ/graph.json","fetch_events":"https://pith.science/api/pith-number/NOKRD3ZLF7WJFJ7XS2SO5HRQRJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NOKRD3ZLF7WJFJ7XS2SO5HRQRJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NOKRD3ZLF7WJFJ7XS2SO5HRQRJ/action/storage_attestation","attest_author":"https://pith.science/pith/NOKRD3ZLF7WJFJ7XS2SO5HRQRJ/action/author_attestation","sign_citation":"https://pith.science/pith/NOKRD3ZLF7WJFJ7XS2SO5HRQRJ/action/citation_signature","submit_replication":"https://pith.science/pith/NOKRD3ZLF7WJFJ7XS2SO5HRQRJ/action/replication_record"}},"created_at":"2026-07-05T11:34:04.347123+00:00","updated_at":"2026-07-05T11:34:04.347123+00:00"}