{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2015:IVQEY3YUYJZL2EHSQA6L3FW6SJ","short_pith_number":"pith:IVQEY3YU","schema_version":"1.0","canonical_sha256":"45604c6f14c272bd10f2803cbd96de926b87a9f7c1f66482425cc4a82f8d5a2f","source":{"kind":"arxiv","id":"1502.05477","version":5},"attestation_state":"computed","paper":{"title":"Trust Region Policy Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"John Schulman, Michael I. Jordan, Philipp Moritz, Pieter Abbeel, Sergey Levine","submitted_at":"2015-02-19T06:44:25Z","abstract_excerpt":"We describe an iterative procedure for optimizing policies, with guaranteed monotonic improvement. By making several approximations to the theoretically-justified procedure, we develop a practical algorithm, called Trust Region Policy Optimization (TRPO). This algorithm is similar to natural policy gradient methods and is effective for optimizing large nonlinear policies such as neural networks. Our experiments demonstrate its robust performance on a wide variety of tasks: learning simulated robotic swimming, hopping, and walking gaits; and playing Atari games using images of the screen as inp"},"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":"1502.05477","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2015-02-19T06:44:25Z","cross_cats_sorted":[],"title_canon_sha256":"b60dc0f59dcb0084c59dd62e514ae98c5367cb138d520229a2682bc708634d40","abstract_canon_sha256":"b40778b0e01a0bac6fc7f67f1bb23ef7affec61a5861fec6cbf8679733aa488f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:46:02.905473Z","signature_b64":"2tQ6bbFP+b2d4mPFeM6ayWN88Oi1cQrcw/yvrXYfue603cX6ti4SARHPfYx0YYiMX3iY+l3FCWV8l+7xdVH1Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"45604c6f14c272bd10f2803cbd96de926b87a9f7c1f66482425cc4a82f8d5a2f","last_reissued_at":"2026-05-18T00:46:02.904977Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:46:02.904977Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Trust Region Policy Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"John Schulman, Michael I. Jordan, Philipp Moritz, Pieter Abbeel, Sergey Levine","submitted_at":"2015-02-19T06:44:25Z","abstract_excerpt":"We describe an iterative procedure for optimizing policies, with guaranteed monotonic improvement. By making several approximations to the theoretically-justified procedure, we develop a practical algorithm, called Trust Region Policy Optimization (TRPO). This algorithm is similar to natural policy gradient methods and is effective for optimizing large nonlinear policies such as neural networks. Our experiments demonstrate its robust performance on a wide variety of tasks: learning simulated robotic swimming, hopping, and walking gaits; and playing Atari games using images of the screen as inp"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1502.05477","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":""},"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":"1502.05477","created_at":"2026-05-18T00:46:02.905056+00:00"},{"alias_kind":"arxiv_version","alias_value":"1502.05477v5","created_at":"2026-05-18T00:46:02.905056+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1502.05477","created_at":"2026-05-18T00:46:02.905056+00:00"},{"alias_kind":"pith_short_12","alias_value":"IVQEY3YUYJZL","created_at":"2026-05-18T12:29:25.134429+00:00"},{"alias_kind":"pith_short_16","alias_value":"IVQEY3YUYJZL2EHS","created_at":"2026-05-18T12:29:25.134429+00:00"},{"alias_kind":"pith_short_8","alias_value":"IVQEY3YU","created_at":"2026-05-18T12:29:25.134429+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":20,"internal_anchor_count":11,"sample":[{"citing_arxiv_id":"2312.09436","citing_title":"Temporal Transfer Learning for Traffic Optimization with Coarse-grained Advisory Autonomy","ref_index":61,"is_internal_anchor":true},{"citing_arxiv_id":"2408.09468","citing_title":"Towards Safe and Robust Autonomous Vehicle Platooning: A Self-Organizing Cooperative Control Framework","ref_index":23,"is_internal_anchor":true},{"citing_arxiv_id":"2410.06347","citing_title":"Goal-Conditioned Decision Transformer for Multi-Goal Offline Reinforcement Learning","ref_index":36,"is_internal_anchor":true},{"citing_arxiv_id":"2504.06355","citing_title":"An Information-Geometric Approach to Artificial Curiosity","ref_index":11,"is_internal_anchor":true},{"citing_arxiv_id":"2508.16474","citing_title":"Reinforcement Learning-based Control via Y-wise Affine Neural Networks (YANNs)","ref_index":63,"is_internal_anchor":true},{"citing_arxiv_id":"2509.22963","citing_title":"Reinforcement Learning with Discrete Diffusion Policies for Combinatorial Action Spaces","ref_index":41,"is_internal_anchor":true},{"citing_arxiv_id":"2605.19425","citing_title":"When to Stop Reusing: Dynamic Gradient Gating for Sample-Efficient RLVR","ref_index":20,"is_internal_anchor":true},{"citing_arxiv_id":"2509.09838","citing_title":"Dissecting Discrete Soft Actor-Critic: Limitations and Principled Alternatives","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2509.25424","citing_title":"Polychromic Objectives for Reinforcement Learning","ref_index":30,"is_internal_anchor":true},{"citing_arxiv_id":"2602.19837","citing_title":"Meta-Learning and Meta-Reinforcement Learning -- Tracing the Path towards DeepMind's Adaptive Agent","ref_index":141,"is_internal_anchor":true},{"citing_arxiv_id":"2605.14366","citing_title":"Reinforcement Learning with Semantic Rewards Enables Low-Resource Language Expansion without Alignment Tax","ref_index":11,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08378","citing_title":"Reinforcement Learning for Scalable and Trustworthy Intelligent Systems","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09214","citing_title":"Fast Rates for Offline Contextual Bandits with Forward-KL Regularization under Single-Policy Concentrability","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2604.23716","citing_title":"Information-Theoretic Measures in AI: A Practical Decision Guide","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"1509.02971","citing_title":"Continuous control with deep reinforcement learning","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2604.09035","citing_title":"Advantage-Guided Diffusion for Model-Based Reinforcement Learning","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"1506.02438","citing_title":"High-Dimensional Continuous Control Using Generalized Advantage Estimation","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05394","citing_title":"Neural Assistive Impulses: Synthesizing Exaggerated Motions for Physics-based Characters","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2604.17706","citing_title":"OmniVLA-RL: A Vision-Language-Action Model with Spatial Understanding and Online RL","ref_index":48,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19695","citing_title":"Planning in entropy-regularized Markov decision processes and games","ref_index":22,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IVQEY3YUYJZL2EHSQA6L3FW6SJ","json":"https://pith.science/pith/IVQEY3YUYJZL2EHSQA6L3FW6SJ.json","graph_json":"https://pith.science/api/pith-number/IVQEY3YUYJZL2EHSQA6L3FW6SJ/graph.json","events_json":"https://pith.science/api/pith-number/IVQEY3YUYJZL2EHSQA6L3FW6SJ/events.json","paper":"https://pith.science/paper/IVQEY3YU"},"agent_actions":{"view_html":"https://pith.science/pith/IVQEY3YUYJZL2EHSQA6L3FW6SJ","download_json":"https://pith.science/pith/IVQEY3YUYJZL2EHSQA6L3FW6SJ.json","view_paper":"https://pith.science/paper/IVQEY3YU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1502.05477&json=true","fetch_graph":"https://pith.science/api/pith-number/IVQEY3YUYJZL2EHSQA6L3FW6SJ/graph.json","fetch_events":"https://pith.science/api/pith-number/IVQEY3YUYJZL2EHSQA6L3FW6SJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IVQEY3YUYJZL2EHSQA6L3FW6SJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IVQEY3YUYJZL2EHSQA6L3FW6SJ/action/storage_attestation","attest_author":"https://pith.science/pith/IVQEY3YUYJZL2EHSQA6L3FW6SJ/action/author_attestation","sign_citation":"https://pith.science/pith/IVQEY3YUYJZL2EHSQA6L3FW6SJ/action/citation_signature","submit_replication":"https://pith.science/pith/IVQEY3YUYJZL2EHSQA6L3FW6SJ/action/replication_record"}},"created_at":"2026-05-18T00:46:02.905056+00:00","updated_at":"2026-05-18T00:46:02.905056+00:00"}