{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:NHW6WUC67VMTQ4AALUDVMDTRBJ","short_pith_number":"pith:NHW6WUC6","schema_version":"1.0","canonical_sha256":"69edeb505efd593870005d07560e710a4df7a2a2695d6c51652b04fc018a60b7","source":{"kind":"arxiv","id":"2506.05615","version":1},"attestation_state":"computed","paper":{"title":"When Maximum Entropy Misleads Policy Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Ruipeng Zhang, Sicun Gao, Ya-Chien Chang","submitted_at":"2025-06-05T22:03:38Z","abstract_excerpt":"The Maximum Entropy Reinforcement Learning (MaxEnt RL) framework is a leading approach for achieving efficient learning and robust performance across many RL tasks. However, MaxEnt methods have also been shown to struggle with performance-critical control problems in practice, where non-MaxEnt algorithms can successfully learn. In this work, we analyze how the trade-off between robustness and optimality affects the performance of MaxEnt algorithms in complex control tasks: while entropy maximization enhances exploration and robustness, it can also mislead policy optimization, leading to failur"},"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":"2506.05615","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-06-05T22:03:38Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"4ade1576e8674c4579547e633d5bd3c0d418986b4cd487ca62c58dcaece502d5","abstract_canon_sha256":"4a1cd0c878f46d4c79886d30788b7e13e6271fa827688f989663ec6d470bb22a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:20:13.203241Z","signature_b64":"POT1o3uJXjGJA7J+GLkACnn0nn0tOfG/lKP1ImNaDMcPQkT6TR+O6g+FJ7Zs86USCzYcfGWD5g5sMdVC0eP5DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"69edeb505efd593870005d07560e710a4df7a2a2695d6c51652b04fc018a60b7","last_reissued_at":"2026-07-05T11:20:13.202737Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:20:13.202737Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"When Maximum Entropy Misleads Policy Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Ruipeng Zhang, Sicun Gao, Ya-Chien Chang","submitted_at":"2025-06-05T22:03:38Z","abstract_excerpt":"The Maximum Entropy Reinforcement Learning (MaxEnt RL) framework is a leading approach for achieving efficient learning and robust performance across many RL tasks. However, MaxEnt methods have also been shown to struggle with performance-critical control problems in practice, where non-MaxEnt algorithms can successfully learn. In this work, we analyze how the trade-off between robustness and optimality affects the performance of MaxEnt algorithms in complex control tasks: while entropy maximization enhances exploration and robustness, it can also mislead policy optimization, leading to failur"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.05615","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/2506.05615/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":"2506.05615","created_at":"2026-07-05T11:20:13.202798+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.05615v1","created_at":"2026-07-05T11:20:13.202798+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.05615","created_at":"2026-07-05T11:20:13.202798+00:00"},{"alias_kind":"pith_short_12","alias_value":"NHW6WUC67VMT","created_at":"2026-07-05T11:20:13.202798+00:00"},{"alias_kind":"pith_short_16","alias_value":"NHW6WUC67VMTQ4AA","created_at":"2026-07-05T11:20:13.202798+00:00"},{"alias_kind":"pith_short_8","alias_value":"NHW6WUC6","created_at":"2026-07-05T11:20:13.202798+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2509.20265","citing_title":"Failure Modes of Maximum Entropy RLHF","ref_index":59,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11491","citing_title":"Understanding and Preventing Entropy Collapse in RLVR with On-Policy Entropy Flow Optimization","ref_index":43,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NHW6WUC67VMTQ4AALUDVMDTRBJ","json":"https://pith.science/pith/NHW6WUC67VMTQ4AALUDVMDTRBJ.json","graph_json":"https://pith.science/api/pith-number/NHW6WUC67VMTQ4AALUDVMDTRBJ/graph.json","events_json":"https://pith.science/api/pith-number/NHW6WUC67VMTQ4AALUDVMDTRBJ/events.json","paper":"https://pith.science/paper/NHW6WUC6"},"agent_actions":{"view_html":"https://pith.science/pith/NHW6WUC67VMTQ4AALUDVMDTRBJ","download_json":"https://pith.science/pith/NHW6WUC67VMTQ4AALUDVMDTRBJ.json","view_paper":"https://pith.science/paper/NHW6WUC6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.05615&json=true","fetch_graph":"https://pith.science/api/pith-number/NHW6WUC67VMTQ4AALUDVMDTRBJ/graph.json","fetch_events":"https://pith.science/api/pith-number/NHW6WUC67VMTQ4AALUDVMDTRBJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NHW6WUC67VMTQ4AALUDVMDTRBJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NHW6WUC67VMTQ4AALUDVMDTRBJ/action/storage_attestation","attest_author":"https://pith.science/pith/NHW6WUC67VMTQ4AALUDVMDTRBJ/action/author_attestation","sign_citation":"https://pith.science/pith/NHW6WUC67VMTQ4AALUDVMDTRBJ/action/citation_signature","submit_replication":"https://pith.science/pith/NHW6WUC67VMTQ4AALUDVMDTRBJ/action/replication_record"}},"created_at":"2026-07-05T11:20:13.202798+00:00","updated_at":"2026-07-05T11:20:13.202798+00:00"}