{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:WHFQHZY5J3RQSMIQH4RPCDED5D","short_pith_number":"pith:WHFQHZY5","schema_version":"1.0","canonical_sha256":"b1cb03e71d4ee30931103f22f10c83e8d2fd179f748c5504aeab3ff52ccc1bad","source":{"kind":"arxiv","id":"2501.15556","version":1},"attestation_state":"computed","paper":{"title":"Commute Your Domains: Trajectory Optimality Criterion for Multi-Domain Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Alexander Podolskiy, Alexey Rukhovich, Irina Piontkovskaya","submitted_at":"2025-01-26T15:12:06Z","abstract_excerpt":"In multi-domain learning, a single model is trained on diverse data domains to leverage shared knowledge and improve generalization. The order in which the data from these domains is used for training can significantly affect the model's performance on each domain. However, this dependence is under-studied. In this paper, we investigate the influence of training order (or data mixing) in multi-domain learning using the concept of Lie bracket of gradient vector fields. By analyzing the infinitesimal effects of changing the training order, we identify regions in the parameter space where alterin"},"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":"2501.15556","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-01-26T15:12:06Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"4b272e5a03f66821943b58c796161b8cf8de7ce198354d49cd8d031ff6d30764","abstract_canon_sha256":"b699583aa6a911978b7c7ff77d6ce06eeb1aa5f4f4f700a2bb5507ffd1191bc2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:05:31.030414Z","signature_b64":"GKcuY7iQ2Z6ZT4NkVIaK+58vPIUuSZBg2+UwBzEd5VdEcxuZBKWZaNDvnb4X3xItDDFQkFMFkDlZLI1lENbgAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b1cb03e71d4ee30931103f22f10c83e8d2fd179f748c5504aeab3ff52ccc1bad","last_reissued_at":"2026-07-05T10:05:31.029903Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:05:31.029903Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Commute Your Domains: Trajectory Optimality Criterion for Multi-Domain Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Alexander Podolskiy, Alexey Rukhovich, Irina Piontkovskaya","submitted_at":"2025-01-26T15:12:06Z","abstract_excerpt":"In multi-domain learning, a single model is trained on diverse data domains to leverage shared knowledge and improve generalization. The order in which the data from these domains is used for training can significantly affect the model's performance on each domain. However, this dependence is under-studied. In this paper, we investigate the influence of training order (or data mixing) in multi-domain learning using the concept of Lie bracket of gradient vector fields. By analyzing the infinitesimal effects of changing the training order, we identify regions in the parameter space where alterin"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.15556","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/2501.15556/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":"2501.15556","created_at":"2026-07-05T10:05:31.029966+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.15556v1","created_at":"2026-07-05T10:05:31.029966+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.15556","created_at":"2026-07-05T10:05:31.029966+00:00"},{"alias_kind":"pith_short_12","alias_value":"WHFQHZY5J3RQ","created_at":"2026-07-05T10:05:31.029966+00:00"},{"alias_kind":"pith_short_16","alias_value":"WHFQHZY5J3RQSMIQ","created_at":"2026-07-05T10:05:31.029966+00:00"},{"alias_kind":"pith_short_8","alias_value":"WHFQHZY5","created_at":"2026-07-05T10:05:31.029966+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.24993","citing_title":"The Geometry of Sequential Learning: Lie-Bracket Prediction of Transfer Order","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29554","citing_title":"Optimizer Memory Makes Shuffle Order a First-Order Source of Fine-Tuning Noise","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2507.15640","citing_title":"Data Mixing Agent: Learning to Re-weight Domains for Continual Pre-training","ref_index":26,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WHFQHZY5J3RQSMIQH4RPCDED5D","json":"https://pith.science/pith/WHFQHZY5J3RQSMIQH4RPCDED5D.json","graph_json":"https://pith.science/api/pith-number/WHFQHZY5J3RQSMIQH4RPCDED5D/graph.json","events_json":"https://pith.science/api/pith-number/WHFQHZY5J3RQSMIQH4RPCDED5D/events.json","paper":"https://pith.science/paper/WHFQHZY5"},"agent_actions":{"view_html":"https://pith.science/pith/WHFQHZY5J3RQSMIQH4RPCDED5D","download_json":"https://pith.science/pith/WHFQHZY5J3RQSMIQH4RPCDED5D.json","view_paper":"https://pith.science/paper/WHFQHZY5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.15556&json=true","fetch_graph":"https://pith.science/api/pith-number/WHFQHZY5J3RQSMIQH4RPCDED5D/graph.json","fetch_events":"https://pith.science/api/pith-number/WHFQHZY5J3RQSMIQH4RPCDED5D/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WHFQHZY5J3RQSMIQH4RPCDED5D/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WHFQHZY5J3RQSMIQH4RPCDED5D/action/storage_attestation","attest_author":"https://pith.science/pith/WHFQHZY5J3RQSMIQH4RPCDED5D/action/author_attestation","sign_citation":"https://pith.science/pith/WHFQHZY5J3RQSMIQH4RPCDED5D/action/citation_signature","submit_replication":"https://pith.science/pith/WHFQHZY5J3RQSMIQH4RPCDED5D/action/replication_record"}},"created_at":"2026-07-05T10:05:31.029966+00:00","updated_at":"2026-07-05T10:05:31.029966+00:00"}