{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:RW2J74PZQYVRRTSOOAC7P4IKCY","short_pith_number":"pith:RW2J74PZ","schema_version":"1.0","canonical_sha256":"8db49ff1f9862b18ce4e7005f7f10a16167383e0375df560e248daab3683ef9a","source":{"kind":"arxiv","id":"2605.20341","version":1},"attestation_state":"computed","paper":{"title":"Causal Unlearning in Collaborative Optimization: Exact and Approximate Influence Reversal under Adversarial Contributions","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CR","cs.PF"],"primary_cat":"cs.LG","authors_text":"Ali Mahdavi, Amirfarhad Farhadi, Azadeh Zamanifar, Omid Kashefi","submitted_at":"2026-05-19T18:00:39Z","abstract_excerpt":"Federated learning systems must support data deletion requests to comply with privacy regulations, yet retraining from scratch after each deletion is computationally prohibitive. We present HF-KCU, a method that removes a client's contribution by approximating the influence function through conjugate gradient iterations in Krylov subspaces, reducing complexity from O(d^3) to O(kd) where k<<d.A causal weighting mechanism ensures that only clients holding the deleted data receive parameter updates, preventing spurious changes to unaffected clients. Our method is designed to handle bounded advers"},"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":"2605.20341","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-19T18:00:39Z","cross_cats_sorted":["cs.AI","cs.CR","cs.PF"],"title_canon_sha256":"9d0855588e84d42f7ae95d7dcba0af0929e1878e43a35b65e0fb011e0b60b895","abstract_canon_sha256":"a1b68352dfe0019033b5769cea1d083a56d5b16000142499fc4a78a934f88bd1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T00:04:25.932514Z","signature_b64":"I+woE5/M7jQUWy7YKl5KXYU+xRsN+9LK0Zwb7yv2SH9DJ2fnEREku0jH5xsq06JIHCG0OToLBNOa8psqv8/oBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8db49ff1f9862b18ce4e7005f7f10a16167383e0375df560e248daab3683ef9a","last_reissued_at":"2026-05-21T00:04:25.932053Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T00:04:25.932053Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Causal Unlearning in Collaborative Optimization: Exact and Approximate Influence Reversal under Adversarial Contributions","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CR","cs.PF"],"primary_cat":"cs.LG","authors_text":"Ali Mahdavi, Amirfarhad Farhadi, Azadeh Zamanifar, Omid Kashefi","submitted_at":"2026-05-19T18:00:39Z","abstract_excerpt":"Federated learning systems must support data deletion requests to comply with privacy regulations, yet retraining from scratch after each deletion is computationally prohibitive. We present HF-KCU, a method that removes a client's contribution by approximating the influence function through conjugate gradient iterations in Krylov subspaces, reducing complexity from O(d^3) to O(kd) where k<<d.A causal weighting mechanism ensures that only clients holding the deleted data receive parameter updates, preventing spurious changes to unaffected clients. Our method is designed to handle bounded advers"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.20341","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/2605.20341/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":"2605.20341","created_at":"2026-05-21T00:04:25.932114+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.20341v1","created_at":"2026-05-21T00:04:25.932114+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.20341","created_at":"2026-05-21T00:04:25.932114+00:00"},{"alias_kind":"pith_short_12","alias_value":"RW2J74PZQYVR","created_at":"2026-05-21T00:04:25.932114+00:00"},{"alias_kind":"pith_short_16","alias_value":"RW2J74PZQYVRRTSO","created_at":"2026-05-21T00:04:25.932114+00:00"},{"alias_kind":"pith_short_8","alias_value":"RW2J74PZ","created_at":"2026-05-21T00:04:25.932114+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/RW2J74PZQYVRRTSOOAC7P4IKCY","json":"https://pith.science/pith/RW2J74PZQYVRRTSOOAC7P4IKCY.json","graph_json":"https://pith.science/api/pith-number/RW2J74PZQYVRRTSOOAC7P4IKCY/graph.json","events_json":"https://pith.science/api/pith-number/RW2J74PZQYVRRTSOOAC7P4IKCY/events.json","paper":"https://pith.science/paper/RW2J74PZ"},"agent_actions":{"view_html":"https://pith.science/pith/RW2J74PZQYVRRTSOOAC7P4IKCY","download_json":"https://pith.science/pith/RW2J74PZQYVRRTSOOAC7P4IKCY.json","view_paper":"https://pith.science/paper/RW2J74PZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.20341&json=true","fetch_graph":"https://pith.science/api/pith-number/RW2J74PZQYVRRTSOOAC7P4IKCY/graph.json","fetch_events":"https://pith.science/api/pith-number/RW2J74PZQYVRRTSOOAC7P4IKCY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RW2J74PZQYVRRTSOOAC7P4IKCY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RW2J74PZQYVRRTSOOAC7P4IKCY/action/storage_attestation","attest_author":"https://pith.science/pith/RW2J74PZQYVRRTSOOAC7P4IKCY/action/author_attestation","sign_citation":"https://pith.science/pith/RW2J74PZQYVRRTSOOAC7P4IKCY/action/citation_signature","submit_replication":"https://pith.science/pith/RW2J74PZQYVRRTSOOAC7P4IKCY/action/replication_record"}},"created_at":"2026-05-21T00:04:25.932114+00:00","updated_at":"2026-05-21T00:04:25.932114+00:00"}