{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:FIN5YJADZOS4QZCY4EQPFXCW7W","short_pith_number":"pith:FIN5YJAD","schema_version":"1.0","canonical_sha256":"2a1bdc2403cba5c86458e120f2dc56fdb9d68a996c0cfbc8efab3f81b1cd4721","source":{"kind":"arxiv","id":"2005.09814","version":5},"attestation_state":"computed","paper":{"title":"Mirror Descent Policy Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Lior Shani, Manan Tomar, Mohammad Ghavamzadeh, Yonathan Efroni","submitted_at":"2020-05-20T01:30:43Z","abstract_excerpt":"Mirror descent (MD), a well-known first-order method in constrained convex optimization, has recently been shown as an important tool to analyze trust-region algorithms in reinforcement learning (RL). However, there remains a considerable gap between such theoretically analyzed algorithms and the ones used in practice. Inspired by this, we propose an efficient RL algorithm, called {\\em mirror descent policy optimization} (MDPO). MDPO iteratively updates the policy by {\\em approximately} solving a trust-region problem, whose objective function consists of two terms: a linearization of the stand"},"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":"2005.09814","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-05-20T01:30:43Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"becb9f897a3403f9b5324b46cab49a738110d689afe24c09f59c13349b4caa7e","abstract_canon_sha256":"07b6ecdf80008c2eca0e3fe913cdf1bc9b18f9a89911f50a1d0f9f2643bb0416"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:46:40.750833Z","signature_b64":"ocLjLIDbc5XRA1Je/NCYnQHmvrR0haCKnEik8bV8pwTNozToqrhitjDeH7kHK5OkQQpsjC52f46nrq3D70SbCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2a1bdc2403cba5c86458e120f2dc56fdb9d68a996c0cfbc8efab3f81b1cd4721","last_reissued_at":"2026-07-05T02:46:40.750341Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:46:40.750341Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Mirror Descent Policy Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Lior Shani, Manan Tomar, Mohammad Ghavamzadeh, Yonathan Efroni","submitted_at":"2020-05-20T01:30:43Z","abstract_excerpt":"Mirror descent (MD), a well-known first-order method in constrained convex optimization, has recently been shown as an important tool to analyze trust-region algorithms in reinforcement learning (RL). However, there remains a considerable gap between such theoretically analyzed algorithms and the ones used in practice. Inspired by this, we propose an efficient RL algorithm, called {\\em mirror descent policy optimization} (MDPO). MDPO iteratively updates the policy by {\\em approximately} solving a trust-region problem, whose objective function consists of two terms: a linearization of the stand"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2005.09814","kind":"arxiv","version":5},"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/2005.09814/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":"2005.09814","created_at":"2026-07-05T02:46:40.750400+00:00"},{"alias_kind":"arxiv_version","alias_value":"2005.09814v5","created_at":"2026-07-05T02:46:40.750400+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2005.09814","created_at":"2026-07-05T02:46:40.750400+00:00"},{"alias_kind":"pith_short_12","alias_value":"FIN5YJADZOS4","created_at":"2026-07-05T02:46:40.750400+00:00"},{"alias_kind":"pith_short_16","alias_value":"FIN5YJADZOS4QZCY","created_at":"2026-07-05T02:46:40.750400+00:00"},{"alias_kind":"pith_short_8","alias_value":"FIN5YJAD","created_at":"2026-07-05T02:46:40.750400+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":13,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.29092","citing_title":"Priced Motion Through Optimal Faces: A Normal-Fan Geometry for Non-Stationary Adversarial MDPs","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29092","citing_title":"Priced Motion Through Optimal Faces: A Normal-Fan Geometry for Non-Stationary Adversarial MDPs","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25507","citing_title":"Credit Assignment with Resets in Language Model Reasoning","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2402.08106","citing_title":"Mirror Descent-Ascent for mean-field min-max problems","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2509.22963","citing_title":"Reinforcement Learning with Discrete Diffusion Policies for Combinatorial Action Spaces","ref_index":55,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06139","citing_title":"Listwise Policy Optimization: Group-based RLVR as Target-Projection on the LLM Response Simplex","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18591","citing_title":"Randomized Advantage Transformation (RAT): Computing Natural Policy Gradients via Direct Backpropagation","ref_index":137,"is_internal_anchor":false},{"citing_arxiv_id":"2509.09838","citing_title":"Dissecting Discrete Soft Actor-Critic: Limitations and Principled Alternatives","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2503.16419","citing_title":"Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models","ref_index":173,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03921","citing_title":"Optimal Posterior Sampling for Policy Identification in Tabular Markov Decision Processes","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2502.16982","citing_title":"Muon is Scalable for LLM Training","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06165","citing_title":"Post Reasoning: Improving the Performance of Non-Thinking Models at No Cost","ref_index":95,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06139","citing_title":"Listwise Policy Optimization: Group-based RLVR as Target-Projection on the LLM Response Simplex","ref_index":22,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FIN5YJADZOS4QZCY4EQPFXCW7W","json":"https://pith.science/pith/FIN5YJADZOS4QZCY4EQPFXCW7W.json","graph_json":"https://pith.science/api/pith-number/FIN5YJADZOS4QZCY4EQPFXCW7W/graph.json","events_json":"https://pith.science/api/pith-number/FIN5YJADZOS4QZCY4EQPFXCW7W/events.json","paper":"https://pith.science/paper/FIN5YJAD"},"agent_actions":{"view_html":"https://pith.science/pith/FIN5YJADZOS4QZCY4EQPFXCW7W","download_json":"https://pith.science/pith/FIN5YJADZOS4QZCY4EQPFXCW7W.json","view_paper":"https://pith.science/paper/FIN5YJAD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2005.09814&json=true","fetch_graph":"https://pith.science/api/pith-number/FIN5YJADZOS4QZCY4EQPFXCW7W/graph.json","fetch_events":"https://pith.science/api/pith-number/FIN5YJADZOS4QZCY4EQPFXCW7W/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FIN5YJADZOS4QZCY4EQPFXCW7W/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FIN5YJADZOS4QZCY4EQPFXCW7W/action/storage_attestation","attest_author":"https://pith.science/pith/FIN5YJADZOS4QZCY4EQPFXCW7W/action/author_attestation","sign_citation":"https://pith.science/pith/FIN5YJADZOS4QZCY4EQPFXCW7W/action/citation_signature","submit_replication":"https://pith.science/pith/FIN5YJADZOS4QZCY4EQPFXCW7W/action/replication_record"}},"created_at":"2026-07-05T02:46:40.750400+00:00","updated_at":"2026-07-05T02:46:40.750400+00:00"}