{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:PPZQ3XFU4YK536VBV7JLBQX5VA","short_pith_number":"pith:PPZQ3XFU","schema_version":"1.0","canonical_sha256":"7bf30ddcb4e615ddfaa1afd2b0c2fda8313ecd958c95551c3aa66c7f364c0ead","source":{"kind":"arxiv","id":"2003.04108","version":2},"attestation_state":"computed","paper":{"title":"Stable Policy Optimization via Off-Policy Divergence Regularization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Ahmed Touati, Amy Zhang, Joelle Pineau, Pascal Vincent","submitted_at":"2020-03-09T13:05:47Z","abstract_excerpt":"Trust Region Policy Optimization (TRPO) and Proximal Policy Optimization (PPO) are among the most successful policy gradient approaches in deep reinforcement learning (RL). While these methods achieve state-of-the-art performance across a wide range of challenging tasks, there is room for improvement in the stabilization of the policy learning and how the off-policy data are used. In this paper we revisit the theoretical foundations of these algorithms and propose a new algorithm which stabilizes the policy improvement through a proximity term that constrains the discounted state-action visita"},"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":"2003.04108","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-03-09T13:05:47Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"e4899f4f6dad38f47f139b14ca1b3aa566209cb5a4f01fdfe2e5f4ee3c9fea01","abstract_canon_sha256":"cd6edb21a9ae499b45672a5d624f174422555f4123aff9f5c1f8572010e13d49"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:11:34.121102Z","signature_b64":"ZDL/z9O9ogkFipK2PqgHsXeKgGwD/V+p0WHDS/f+/TqGcugxJcRPGUaknpjTYXIg8GSHK63HlRRoTvsBSLdiCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7bf30ddcb4e615ddfaa1afd2b0c2fda8313ecd958c95551c3aa66c7f364c0ead","last_reissued_at":"2026-07-05T01:11:34.120666Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:11:34.120666Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Stable Policy Optimization via Off-Policy Divergence Regularization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Ahmed Touati, Amy Zhang, Joelle Pineau, Pascal Vincent","submitted_at":"2020-03-09T13:05:47Z","abstract_excerpt":"Trust Region Policy Optimization (TRPO) and Proximal Policy Optimization (PPO) are among the most successful policy gradient approaches in deep reinforcement learning (RL). While these methods achieve state-of-the-art performance across a wide range of challenging tasks, there is room for improvement in the stabilization of the policy learning and how the off-policy data are used. In this paper we revisit the theoretical foundations of these algorithms and propose a new algorithm which stabilizes the policy improvement through a proximity term that constrains the discounted state-action visita"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2003.04108","kind":"arxiv","version":2},"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/2003.04108/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":"2003.04108","created_at":"2026-07-05T01:11:34.120730+00:00"},{"alias_kind":"arxiv_version","alias_value":"2003.04108v2","created_at":"2026-07-05T01:11:34.120730+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2003.04108","created_at":"2026-07-05T01:11:34.120730+00:00"},{"alias_kind":"pith_short_12","alias_value":"PPZQ3XFU4YK5","created_at":"2026-07-05T01:11:34.120730+00:00"},{"alias_kind":"pith_short_16","alias_value":"PPZQ3XFU4YK536VB","created_at":"2026-07-05T01:11:34.120730+00:00"},{"alias_kind":"pith_short_8","alias_value":"PPZQ3XFU","created_at":"2026-07-05T01:11:34.120730+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/PPZQ3XFU4YK536VBV7JLBQX5VA","json":"https://pith.science/pith/PPZQ3XFU4YK536VBV7JLBQX5VA.json","graph_json":"https://pith.science/api/pith-number/PPZQ3XFU4YK536VBV7JLBQX5VA/graph.json","events_json":"https://pith.science/api/pith-number/PPZQ3XFU4YK536VBV7JLBQX5VA/events.json","paper":"https://pith.science/paper/PPZQ3XFU"},"agent_actions":{"view_html":"https://pith.science/pith/PPZQ3XFU4YK536VBV7JLBQX5VA","download_json":"https://pith.science/pith/PPZQ3XFU4YK536VBV7JLBQX5VA.json","view_paper":"https://pith.science/paper/PPZQ3XFU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2003.04108&json=true","fetch_graph":"https://pith.science/api/pith-number/PPZQ3XFU4YK536VBV7JLBQX5VA/graph.json","fetch_events":"https://pith.science/api/pith-number/PPZQ3XFU4YK536VBV7JLBQX5VA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PPZQ3XFU4YK536VBV7JLBQX5VA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PPZQ3XFU4YK536VBV7JLBQX5VA/action/storage_attestation","attest_author":"https://pith.science/pith/PPZQ3XFU4YK536VBV7JLBQX5VA/action/author_attestation","sign_citation":"https://pith.science/pith/PPZQ3XFU4YK536VBV7JLBQX5VA/action/citation_signature","submit_replication":"https://pith.science/pith/PPZQ3XFU4YK536VBV7JLBQX5VA/action/replication_record"}},"created_at":"2026-07-05T01:11:34.120730+00:00","updated_at":"2026-07-05T01:11:34.120730+00:00"}