{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:KG5KDROXQNRTM3CFLQJDICSFHI","short_pith_number":"pith:KG5KDROX","schema_version":"1.0","canonical_sha256":"51baa1c5d78363366c455c12340a453a0ee90e5031d3d4b4b8df3cccb5de12c8","source":{"kind":"arxiv","id":"2502.01145","version":2},"attestation_state":"computed","paper":{"title":"Tackling Feature and Sample Heterogeneity in Decentralized Multi-Task Learning: A Sheaf-Theoretic Approach","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chaouki Ben Issaid, Mehdi Bennis, Praneeth Vepakomma","submitted_at":"2025-02-03T08:25:34Z","abstract_excerpt":"Federated multi-task learning (FMTL) aims to simultaneously learn multiple related tasks across clients without sharing sensitive raw data. However, in the decentralized setting, existing FMTL frameworks are limited in their ability to capture complex task relationships and handle feature and sample heterogeneity across clients. To address these challenges, we introduce a novel sheaf-theoretic-based approach for FMTL. By representing client relationships using cellular sheaves, our framework can flexibly model interactions between heterogeneous client models. We formulate the sheaf-based FMTL "},"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":"2502.01145","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-03T08:25:34Z","cross_cats_sorted":[],"title_canon_sha256":"abc313fa762d7a0bb9c12a338f7f607a560901de167701fe1de75acab2610863","abstract_canon_sha256":"6713a3720a6463708d705a576f39e5bf5df90a70f0aebbdb9a744dfbdab3150a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:17:40.595671Z","signature_b64":"OhRgcoeFvyI4zSUggbD/ldoIBtdUgOFVVnWHMNC2kZC+AhmTSJd1+Z5X7ZPfY+fixhSnA6/8+7as8GmR8dYnBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"51baa1c5d78363366c455c12340a453a0ee90e5031d3d4b4b8df3cccb5de12c8","last_reissued_at":"2026-07-05T11:17:40.594519Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:17:40.594519Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Tackling Feature and Sample Heterogeneity in Decentralized Multi-Task Learning: A Sheaf-Theoretic Approach","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chaouki Ben Issaid, Mehdi Bennis, Praneeth Vepakomma","submitted_at":"2025-02-03T08:25:34Z","abstract_excerpt":"Federated multi-task learning (FMTL) aims to simultaneously learn multiple related tasks across clients without sharing sensitive raw data. However, in the decentralized setting, existing FMTL frameworks are limited in their ability to capture complex task relationships and handle feature and sample heterogeneity across clients. To address these challenges, we introduce a novel sheaf-theoretic-based approach for FMTL. By representing client relationships using cellular sheaves, our framework can flexibly model interactions between heterogeneous client models. We formulate the sheaf-based FMTL "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.01145","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/2502.01145/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":"2502.01145","created_at":"2026-07-05T11:17:40.594627+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.01145v2","created_at":"2026-07-05T11:17:40.594627+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.01145","created_at":"2026-07-05T11:17:40.594627+00:00"},{"alias_kind":"pith_short_12","alias_value":"KG5KDROXQNRT","created_at":"2026-07-05T11:17:40.594627+00:00"},{"alias_kind":"pith_short_16","alias_value":"KG5KDROXQNRTM3CF","created_at":"2026-07-05T11:17:40.594627+00:00"},{"alias_kind":"pith_short_8","alias_value":"KG5KDROX","created_at":"2026-07-05T11:17:40.594627+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.19529","citing_title":"The Sheaf Laplacian: A Topological Framework for Data Fusion and Consensus in Distributed Sensing Networks","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2605.04108","citing_title":"MuCALD-SplitFed: Causal-Latent Diffusion for Privacy-Preserving Multi-Task Split-Federated Medical Image Segmentation","ref_index":10,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KG5KDROXQNRTM3CFLQJDICSFHI","json":"https://pith.science/pith/KG5KDROXQNRTM3CFLQJDICSFHI.json","graph_json":"https://pith.science/api/pith-number/KG5KDROXQNRTM3CFLQJDICSFHI/graph.json","events_json":"https://pith.science/api/pith-number/KG5KDROXQNRTM3CFLQJDICSFHI/events.json","paper":"https://pith.science/paper/KG5KDROX"},"agent_actions":{"view_html":"https://pith.science/pith/KG5KDROXQNRTM3CFLQJDICSFHI","download_json":"https://pith.science/pith/KG5KDROXQNRTM3CFLQJDICSFHI.json","view_paper":"https://pith.science/paper/KG5KDROX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.01145&json=true","fetch_graph":"https://pith.science/api/pith-number/KG5KDROXQNRTM3CFLQJDICSFHI/graph.json","fetch_events":"https://pith.science/api/pith-number/KG5KDROXQNRTM3CFLQJDICSFHI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KG5KDROXQNRTM3CFLQJDICSFHI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KG5KDROXQNRTM3CFLQJDICSFHI/action/storage_attestation","attest_author":"https://pith.science/pith/KG5KDROXQNRTM3CFLQJDICSFHI/action/author_attestation","sign_citation":"https://pith.science/pith/KG5KDROXQNRTM3CFLQJDICSFHI/action/citation_signature","submit_replication":"https://pith.science/pith/KG5KDROXQNRTM3CFLQJDICSFHI/action/replication_record"}},"created_at":"2026-07-05T11:17:40.594627+00:00","updated_at":"2026-07-05T11:17:40.594627+00:00"}