{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:SXPWNL7SVZKTJBRNA2SAUUNLSA","short_pith_number":"pith:SXPWNL7S","schema_version":"1.0","canonical_sha256":"95df66aff2ae5534862d06a40a51ab90312a6e1ea40479ab3f6bf36cb2b8c7af","source":{"kind":"arxiv","id":"2110.13799","version":4},"attestation_state":"computed","paper":{"title":"Neural PPO-Clip Attains Global Optimality: A Hinge Loss Perspective","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hsuan-Yu Yao, I-Chen Wu, Kai-Chun Hu, Kuo-Hao Ho, Liang-Chun Ouyang, Nai-Chieh Huang, Ping-Chun Hsieh","submitted_at":"2021-10-26T15:56:57Z","abstract_excerpt":"Policy optimization is a fundamental principle for designing reinforcement learning algorithms, and one example is the proximal policy optimization algorithm with a clipped surrogate objective (PPO-Clip), which has been popularly used in deep reinforcement learning due to its simplicity and effectiveness. Despite its superior empirical performance, PPO-Clip has not been justified via theoretical proof up to date. In this paper, we establish the first global convergence rate of PPO-Clip under neural function approximation. We identify the fundamental challenges of analyzing PPO-Clip and address"},"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":"2110.13799","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-10-26T15:56:57Z","cross_cats_sorted":[],"title_canon_sha256":"8435f3887a6cf3feb0f24b136f4185a8148d6f077332e67de5064ddf690dfd10","abstract_canon_sha256":"0774f8293eb8a5dd2841b849d0d3379db76089a1e21bd70670ab1d0e91be6772"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:53:39.913578Z","signature_b64":"+nLSTfQx+cydJKPwlWfavYLf4fDf5AqnDXx3y4gIa1/JEhuP6vLQKrXMiUjA23CS8PlEICWcDkmoPosTSh9pDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"95df66aff2ae5534862d06a40a51ab90312a6e1ea40479ab3f6bf36cb2b8c7af","last_reissued_at":"2026-07-05T04:53:39.913172Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:53:39.913172Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Neural PPO-Clip Attains Global Optimality: A Hinge Loss Perspective","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hsuan-Yu Yao, I-Chen Wu, Kai-Chun Hu, Kuo-Hao Ho, Liang-Chun Ouyang, Nai-Chieh Huang, Ping-Chun Hsieh","submitted_at":"2021-10-26T15:56:57Z","abstract_excerpt":"Policy optimization is a fundamental principle for designing reinforcement learning algorithms, and one example is the proximal policy optimization algorithm with a clipped surrogate objective (PPO-Clip), which has been popularly used in deep reinforcement learning due to its simplicity and effectiveness. Despite its superior empirical performance, PPO-Clip has not been justified via theoretical proof up to date. In this paper, we establish the first global convergence rate of PPO-Clip under neural function approximation. We identify the fundamental challenges of analyzing PPO-Clip and address"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2110.13799","kind":"arxiv","version":4},"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/2110.13799/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":"2110.13799","created_at":"2026-07-05T04:53:39.913229+00:00"},{"alias_kind":"arxiv_version","alias_value":"2110.13799v4","created_at":"2026-07-05T04:53:39.913229+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2110.13799","created_at":"2026-07-05T04:53:39.913229+00:00"},{"alias_kind":"pith_short_12","alias_value":"SXPWNL7SVZKT","created_at":"2026-07-05T04:53:39.913229+00:00"},{"alias_kind":"pith_short_16","alias_value":"SXPWNL7SVZKTJBRN","created_at":"2026-07-05T04:53:39.913229+00:00"},{"alias_kind":"pith_short_8","alias_value":"SXPWNL7S","created_at":"2026-07-05T04:53:39.913229+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.25526","citing_title":"Low Variance Trust Region Optimization with Independent Actors and Sequential Updates in Cooperative Multi-agent Reinforcement Learning","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18591","citing_title":"Randomized Advantage Transformation (RAT): Computing Natural Policy Gradients via Direct Backpropagation","ref_index":135,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SXPWNL7SVZKTJBRNA2SAUUNLSA","json":"https://pith.science/pith/SXPWNL7SVZKTJBRNA2SAUUNLSA.json","graph_json":"https://pith.science/api/pith-number/SXPWNL7SVZKTJBRNA2SAUUNLSA/graph.json","events_json":"https://pith.science/api/pith-number/SXPWNL7SVZKTJBRNA2SAUUNLSA/events.json","paper":"https://pith.science/paper/SXPWNL7S"},"agent_actions":{"view_html":"https://pith.science/pith/SXPWNL7SVZKTJBRNA2SAUUNLSA","download_json":"https://pith.science/pith/SXPWNL7SVZKTJBRNA2SAUUNLSA.json","view_paper":"https://pith.science/paper/SXPWNL7S","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2110.13799&json=true","fetch_graph":"https://pith.science/api/pith-number/SXPWNL7SVZKTJBRNA2SAUUNLSA/graph.json","fetch_events":"https://pith.science/api/pith-number/SXPWNL7SVZKTJBRNA2SAUUNLSA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SXPWNL7SVZKTJBRNA2SAUUNLSA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SXPWNL7SVZKTJBRNA2SAUUNLSA/action/storage_attestation","attest_author":"https://pith.science/pith/SXPWNL7SVZKTJBRNA2SAUUNLSA/action/author_attestation","sign_citation":"https://pith.science/pith/SXPWNL7SVZKTJBRNA2SAUUNLSA/action/citation_signature","submit_replication":"https://pith.science/pith/SXPWNL7SVZKTJBRNA2SAUUNLSA/action/replication_record"}},"created_at":"2026-07-05T04:53:39.913229+00:00","updated_at":"2026-07-05T04:53:39.913229+00:00"}