{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:Y5C4OFTEQN3FGYJEOHY42G6KKJ","short_pith_number":"pith:Y5C4OFTE","schema_version":"1.0","canonical_sha256":"c745c71664837653612471f1cd1bca525fe137939e2b36d79a6a6d91cfa519fe","source":{"kind":"arxiv","id":"2405.16077","version":3},"attestation_state":"computed","paper":{"title":"Theoretical Study of Conflict-Avoidant Multi-Objective Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hao Ban, Kaiyi Ji, Peiyao Xiao, Shaofeng Zou, Yudan Wang","submitted_at":"2024-05-25T05:57:46Z","abstract_excerpt":"Multi-task reinforcement learning (MTRL) has shown great promise in many real-world applications. Existing MTRL algorithms often aim to learn a policy that optimizes individual objective functions simultaneously with a given prior preference (or weights) on different tasks. However, these methods often suffer from the issue of \\textit{gradient conflict} such that the tasks with larger gradients dominate the update direction, resulting in a performance degeneration on other tasks. In this paper, we develop a novel dynamic weighting multi-task actor-critic algorithm (MTAC) under two options of s"},"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":"2405.16077","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-25T05:57:46Z","cross_cats_sorted":[],"title_canon_sha256":"914df82fb24e02ccd392b0a1ddbf47cc244a0a5e9744440eb1731789591a4a38","abstract_canon_sha256":"06b514c58b40e56bb5c53a0a1cd8e145e548b43f074e29b588302e2dfe497e5c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:52:38.307987Z","signature_b64":"TStSkF9xKIVzyWi3NdVNQ8DZi3YG7XwmoewymSMo6Ej+2ggxga/OPq6aVJmMbAzEy4TX5VWQ7U9fqVBUnAxCAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c745c71664837653612471f1cd1bca525fe137939e2b36d79a6a6d91cfa519fe","last_reissued_at":"2026-07-05T09:52:38.307486Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:52:38.307486Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Theoretical Study of Conflict-Avoidant Multi-Objective Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hao Ban, Kaiyi Ji, Peiyao Xiao, Shaofeng Zou, Yudan Wang","submitted_at":"2024-05-25T05:57:46Z","abstract_excerpt":"Multi-task reinforcement learning (MTRL) has shown great promise in many real-world applications. Existing MTRL algorithms often aim to learn a policy that optimizes individual objective functions simultaneously with a given prior preference (or weights) on different tasks. However, these methods often suffer from the issue of \\textit{gradient conflict} such that the tasks with larger gradients dominate the update direction, resulting in a performance degeneration on other tasks. In this paper, we develop a novel dynamic weighting multi-task actor-critic algorithm (MTAC) under two options of s"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.16077","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/2405.16077/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":"2405.16077","created_at":"2026-07-05T09:52:38.307551+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.16077v3","created_at":"2026-07-05T09:52:38.307551+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.16077","created_at":"2026-07-05T09:52:38.307551+00:00"},{"alias_kind":"pith_short_12","alias_value":"Y5C4OFTEQN3F","created_at":"2026-07-05T09:52:38.307551+00:00"},{"alias_kind":"pith_short_16","alias_value":"Y5C4OFTEQN3FGYJE","created_at":"2026-07-05T09:52:38.307551+00:00"},{"alias_kind":"pith_short_8","alias_value":"Y5C4OFTE","created_at":"2026-07-05T09:52:38.307551+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.18447","citing_title":"Pessimism Principle Can Be Effective: Towards a Framework for Zero-Shot Transfer Reinforcement Learning","ref_index":85,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Y5C4OFTEQN3FGYJEOHY42G6KKJ","json":"https://pith.science/pith/Y5C4OFTEQN3FGYJEOHY42G6KKJ.json","graph_json":"https://pith.science/api/pith-number/Y5C4OFTEQN3FGYJEOHY42G6KKJ/graph.json","events_json":"https://pith.science/api/pith-number/Y5C4OFTEQN3FGYJEOHY42G6KKJ/events.json","paper":"https://pith.science/paper/Y5C4OFTE"},"agent_actions":{"view_html":"https://pith.science/pith/Y5C4OFTEQN3FGYJEOHY42G6KKJ","download_json":"https://pith.science/pith/Y5C4OFTEQN3FGYJEOHY42G6KKJ.json","view_paper":"https://pith.science/paper/Y5C4OFTE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.16077&json=true","fetch_graph":"https://pith.science/api/pith-number/Y5C4OFTEQN3FGYJEOHY42G6KKJ/graph.json","fetch_events":"https://pith.science/api/pith-number/Y5C4OFTEQN3FGYJEOHY42G6KKJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Y5C4OFTEQN3FGYJEOHY42G6KKJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Y5C4OFTEQN3FGYJEOHY42G6KKJ/action/storage_attestation","attest_author":"https://pith.science/pith/Y5C4OFTEQN3FGYJEOHY42G6KKJ/action/author_attestation","sign_citation":"https://pith.science/pith/Y5C4OFTEQN3FGYJEOHY42G6KKJ/action/citation_signature","submit_replication":"https://pith.science/pith/Y5C4OFTEQN3FGYJEOHY42G6KKJ/action/replication_record"}},"created_at":"2026-07-05T09:52:38.307551+00:00","updated_at":"2026-07-05T09:52:38.307551+00:00"}