{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:ZWIRCKGBG6NTY4V3PYKS3RKKH6","short_pith_number":"pith:ZWIRCKGB","schema_version":"1.0","canonical_sha256":"cd911128c1379b3c72bb7e152dc54a3fbd5ccb05ff6be1d124dd7ccdba8c4a69","source":{"kind":"arxiv","id":"2608.11658","version":1},"attestation_state":"computed","paper":{"title":"Is Per-Agent Policy Composition Safe? Rethinking Successor-Feature Transfer in Cooperative Multi-Agent Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.MA"],"primary_cat":"cs.LG","authors_text":"Sen Li, Zijian Zhao","submitted_at":"2026-08-12T04:56:52Z","abstract_excerpt":"Many reinforcement learning systems, from fleet management to traffic signal control, must serve an objective that changes dynamically after deployment, and retraining a policy for each new objective is prohibitively expensive. For a single agent, this problem is well understood: successor features with generalized policy improvement, together with their universal extension, recombine a library of learned policies into a policy for any new objective, with a guarantee that the result is never worse than any policy in the library. However, multi-agent transfer has received far less attention, an"},"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":"2608.11658","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-08-12T04:56:52Z","cross_cats_sorted":["cs.AI","cs.MA"],"title_canon_sha256":"b78039bb98826321929630f310e8bd840e2d83acd99e6afc82f831a2da1613b0","abstract_canon_sha256":"2af922ba6ad907797b56b9869c5fbe5bd507eabf2403a719609dadca4dc1cd81"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-13T01:27:14.440835Z","signature_b64":"ajTWRDCQjbh94x/h80TDy4xUkUybzZpZWkB6wSOR+5X/y13DmgwmuFhuyrNMz+6Y144Z/6ijhHoPf2LDDj79Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cd911128c1379b3c72bb7e152dc54a3fbd5ccb05ff6be1d124dd7ccdba8c4a69","last_reissued_at":"2026-08-13T01:27:14.438341Z","signature_status":"signed_v1","first_computed_at":"2026-08-13T01:27:14.438341Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Is Per-Agent Policy Composition Safe? Rethinking Successor-Feature Transfer in Cooperative Multi-Agent Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.MA"],"primary_cat":"cs.LG","authors_text":"Sen Li, Zijian Zhao","submitted_at":"2026-08-12T04:56:52Z","abstract_excerpt":"Many reinforcement learning systems, from fleet management to traffic signal control, must serve an objective that changes dynamically after deployment, and retraining a policy for each new objective is prohibitively expensive. For a single agent, this problem is well understood: successor features with generalized policy improvement, together with their universal extension, recombine a library of learned policies into a policy for any new objective, with a guarantee that the result is never worse than any policy in the library. However, multi-agent transfer has received far less attention, an"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.11658","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/2608.11658/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":"2608.11658","created_at":"2026-08-13T01:27:14.439566+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.11658v1","created_at":"2026-08-13T01:27:14.439566+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.11658","created_at":"2026-08-13T01:27:14.439566+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZWIRCKGBG6NT","created_at":"2026-08-13T01:27:14.439566+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZWIRCKGBG6NTY4V3","created_at":"2026-08-13T01:27:14.439566+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZWIRCKGB","created_at":"2026-08-13T01:27:14.439566+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/ZWIRCKGBG6NTY4V3PYKS3RKKH6","json":"https://pith.science/pith/ZWIRCKGBG6NTY4V3PYKS3RKKH6.json","graph_json":"https://pith.science/api/pith-number/ZWIRCKGBG6NTY4V3PYKS3RKKH6/graph.json","events_json":"https://pith.science/api/pith-number/ZWIRCKGBG6NTY4V3PYKS3RKKH6/events.json","paper":"https://pith.science/paper/ZWIRCKGB"},"agent_actions":{"view_html":"https://pith.science/pith/ZWIRCKGBG6NTY4V3PYKS3RKKH6","download_json":"https://pith.science/pith/ZWIRCKGBG6NTY4V3PYKS3RKKH6.json","view_paper":"https://pith.science/paper/ZWIRCKGB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.11658&json=true","fetch_graph":"https://pith.science/api/pith-number/ZWIRCKGBG6NTY4V3PYKS3RKKH6/graph.json","fetch_events":"https://pith.science/api/pith-number/ZWIRCKGBG6NTY4V3PYKS3RKKH6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZWIRCKGBG6NTY4V3PYKS3RKKH6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZWIRCKGBG6NTY4V3PYKS3RKKH6/action/storage_attestation","attest_author":"https://pith.science/pith/ZWIRCKGBG6NTY4V3PYKS3RKKH6/action/author_attestation","sign_citation":"https://pith.science/pith/ZWIRCKGBG6NTY4V3PYKS3RKKH6/action/citation_signature","submit_replication":"https://pith.science/pith/ZWIRCKGBG6NTY4V3PYKS3RKKH6/action/replication_record"}},"created_at":"2026-08-13T01:27:14.439566+00:00","updated_at":"2026-08-13T01:27:14.439566+00:00"}