{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:D7M55WX6PDRZH5G3LF4IKSDPAU","short_pith_number":"pith:D7M55WX6","schema_version":"1.0","canonical_sha256":"1fd9dedafe78e393f4db597885486f050452550146f99834144a46078c8e95d7","source":{"kind":"arxiv","id":"2405.18100","version":3},"attestation_state":"computed","paper":{"title":"A Pontryagin Perspective on Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["math.OC"],"primary_cat":"cs.LG","authors_text":"Claire Vernade, Michael Muehlebach, Onno Eberhard","submitted_at":"2024-05-28T12:05:20Z","abstract_excerpt":"Reinforcement learning has traditionally focused on learning state-dependent policies to solve optimal control problems in a closed-loop fashion. In this work, we introduce the paradigm of open-loop reinforcement learning where a fixed action sequence is learned instead. We present three new algorithms: one robust model-based method and two sample-efficient model-free methods. Rather than basing our algorithms on Bellman's equation from dynamic programming, our work builds on Pontryagin's principle from the theory of open-loop optimal control. We provide convergence guarantees and evaluate all"},"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":"2405.18100","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-28T12:05:20Z","cross_cats_sorted":["math.OC"],"title_canon_sha256":"3943bee0737e4f5b8f63a003895789e05104b1ee6f330a409090a036a738bc11","abstract_canon_sha256":"cefbfb194fb59f9ffb6785953ecd846a9f2a4ad699adbfea6fa8242ed954e722"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:52:12.061431Z","signature_b64":"PzkchlCCTk13VJK88zE3tp5nlycEVvvVG/hvmyHc0PT5ZyRrwR2p1y/+kHMomuiquEJk0mTxuR4AOJt5qel6BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1fd9dedafe78e393f4db597885486f050452550146f99834144a46078c8e95d7","last_reissued_at":"2026-07-05T10:52:12.060942Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:52:12.060942Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Pontryagin Perspective on Reinforcement Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["math.OC"],"primary_cat":"cs.LG","authors_text":"Claire Vernade, Michael Muehlebach, Onno Eberhard","submitted_at":"2024-05-28T12:05:20Z","abstract_excerpt":"Reinforcement learning has traditionally focused on learning state-dependent policies to solve optimal control problems in a closed-loop fashion. In this work, we introduce the paradigm of open-loop reinforcement learning where a fixed action sequence is learned instead. We present three new algorithms: one robust model-based method and two sample-efficient model-free methods. Rather than basing our algorithms on Bellman's equation from dynamic programming, our work builds on Pontryagin's principle from the theory of open-loop optimal control. We provide convergence guarantees and evaluate all"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.18100","kind":"arxiv","version":3},"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/2405.18100/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":"2405.18100","created_at":"2026-07-05T10:52:12.060997+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.18100v3","created_at":"2026-07-05T10:52:12.060997+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.18100","created_at":"2026-07-05T10:52:12.060997+00:00"},{"alias_kind":"pith_short_12","alias_value":"D7M55WX6PDRZ","created_at":"2026-07-05T10:52:12.060997+00:00"},{"alias_kind":"pith_short_16","alias_value":"D7M55WX6PDRZH5G3","created_at":"2026-07-05T10:52:12.060997+00:00"},{"alias_kind":"pith_short_8","alias_value":"D7M55WX6","created_at":"2026-07-05T10:52:12.060997+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.15465","citing_title":"DATA-DRIVEN PRONTO: a Model-free Solution for Numerical Optimal Control","ref_index":14,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/D7M55WX6PDRZH5G3LF4IKSDPAU","json":"https://pith.science/pith/D7M55WX6PDRZH5G3LF4IKSDPAU.json","graph_json":"https://pith.science/api/pith-number/D7M55WX6PDRZH5G3LF4IKSDPAU/graph.json","events_json":"https://pith.science/api/pith-number/D7M55WX6PDRZH5G3LF4IKSDPAU/events.json","paper":"https://pith.science/paper/D7M55WX6"},"agent_actions":{"view_html":"https://pith.science/pith/D7M55WX6PDRZH5G3LF4IKSDPAU","download_json":"https://pith.science/pith/D7M55WX6PDRZH5G3LF4IKSDPAU.json","view_paper":"https://pith.science/paper/D7M55WX6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.18100&json=true","fetch_graph":"https://pith.science/api/pith-number/D7M55WX6PDRZH5G3LF4IKSDPAU/graph.json","fetch_events":"https://pith.science/api/pith-number/D7M55WX6PDRZH5G3LF4IKSDPAU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/D7M55WX6PDRZH5G3LF4IKSDPAU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/D7M55WX6PDRZH5G3LF4IKSDPAU/action/storage_attestation","attest_author":"https://pith.science/pith/D7M55WX6PDRZH5G3LF4IKSDPAU/action/author_attestation","sign_citation":"https://pith.science/pith/D7M55WX6PDRZH5G3LF4IKSDPAU/action/citation_signature","submit_replication":"https://pith.science/pith/D7M55WX6PDRZH5G3LF4IKSDPAU/action/replication_record"}},"created_at":"2026-07-05T10:52:12.060997+00:00","updated_at":"2026-07-05T10:52:12.060997+00:00"}