{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:P2XFTJ5KDN5Z2BAF4J67FLLEHO","short_pith_number":"pith:P2XFTJ5K","schema_version":"1.0","canonical_sha256":"7eae59a7aa1b7b9d0405e27df2ad643bb02409916f5325946a5f798651bca724","source":{"kind":"arxiv","id":"2505.00663","version":1},"attestation_state":"computed","paper":{"title":"Wasserstein Policy Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Brendan Tracey, David Pfau, Diana Borsa, Hado van Hasselt, Ian Davies, Joao G. M. Araujo","submitted_at":"2025-05-01T17:07:01Z","abstract_excerpt":"We introduce Wasserstein Policy Optimization (WPO), an actor-critic algorithm for reinforcement learning in continuous action spaces. WPO can be derived as an approximation to Wasserstein gradient flow over the space of all policies projected into a finite-dimensional parameter space (e.g., the weights of a neural network), leading to a simple and completely general closed-form update. The resulting algorithm combines many properties of deterministic and classic policy gradient methods. Like deterministic policy gradients, it exploits knowledge of the gradient of the action-value function with"},"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":"2505.00663","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-01T17:07:01Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"661baa42e1eb6ea52856990a036b06031c9007b4c7068fdd0167f9357af0c3d8","abstract_canon_sha256":"db68e6b1c90aeeb8e2453ab245a5c6a8da9f4053d930489709da450525dd7eb1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:57:23.677811Z","signature_b64":"AY1m6OJWPjTQ143zgcdHFTfEUdcpjV8094hbL/Kowg6xhFrBH+z+boAaYDhN9OSE0GN5Coh5iZ7XSnbyaqPBBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7eae59a7aa1b7b9d0405e27df2ad643bb02409916f5325946a5f798651bca724","last_reissued_at":"2026-07-05T10:57:23.677331Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:57:23.677331Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Wasserstein Policy Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Brendan Tracey, David Pfau, Diana Borsa, Hado van Hasselt, Ian Davies, Joao G. M. Araujo","submitted_at":"2025-05-01T17:07:01Z","abstract_excerpt":"We introduce Wasserstein Policy Optimization (WPO), an actor-critic algorithm for reinforcement learning in continuous action spaces. WPO can be derived as an approximation to Wasserstein gradient flow over the space of all policies projected into a finite-dimensional parameter space (e.g., the weights of a neural network), leading to a simple and completely general closed-form update. The resulting algorithm combines many properties of deterministic and classic policy gradient methods. Like deterministic policy gradients, it exploits knowledge of the gradient of the action-value function with"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.00663","kind":"arxiv","version":1},"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/2505.00663/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":"2505.00663","created_at":"2026-07-05T10:57:23.677389+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.00663v1","created_at":"2026-07-05T10:57:23.677389+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.00663","created_at":"2026-07-05T10:57:23.677389+00:00"},{"alias_kind":"pith_short_12","alias_value":"P2XFTJ5KDN5Z","created_at":"2026-07-05T10:57:23.677389+00:00"},{"alias_kind":"pith_short_16","alias_value":"P2XFTJ5KDN5Z2BAF","created_at":"2026-07-05T10:57:23.677389+00:00"},{"alias_kind":"pith_short_8","alias_value":"P2XFTJ5K","created_at":"2026-07-05T10:57:23.677389+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.04335","citing_title":"Policy Gradient for Continuous-Time Robust Markov Decision Processes","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2605.26078","citing_title":"Global Convergence of Wasserstein Policy Gradient for Entropy-Regularized Reinforcement Learning","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2605.26784","citing_title":"Ratio-Variance Regularized Policy Optimization","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22622","citing_title":"A note on convergence of Wasserstein policy optimization","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2603.25777","citing_title":"Challenges and opportunities for AI to help deliver fusion energy","ref_index":7,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/P2XFTJ5KDN5Z2BAF4J67FLLEHO","json":"https://pith.science/pith/P2XFTJ5KDN5Z2BAF4J67FLLEHO.json","graph_json":"https://pith.science/api/pith-number/P2XFTJ5KDN5Z2BAF4J67FLLEHO/graph.json","events_json":"https://pith.science/api/pith-number/P2XFTJ5KDN5Z2BAF4J67FLLEHO/events.json","paper":"https://pith.science/paper/P2XFTJ5K"},"agent_actions":{"view_html":"https://pith.science/pith/P2XFTJ5KDN5Z2BAF4J67FLLEHO","download_json":"https://pith.science/pith/P2XFTJ5KDN5Z2BAF4J67FLLEHO.json","view_paper":"https://pith.science/paper/P2XFTJ5K","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.00663&json=true","fetch_graph":"https://pith.science/api/pith-number/P2XFTJ5KDN5Z2BAF4J67FLLEHO/graph.json","fetch_events":"https://pith.science/api/pith-number/P2XFTJ5KDN5Z2BAF4J67FLLEHO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/P2XFTJ5KDN5Z2BAF4J67FLLEHO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/P2XFTJ5KDN5Z2BAF4J67FLLEHO/action/storage_attestation","attest_author":"https://pith.science/pith/P2XFTJ5KDN5Z2BAF4J67FLLEHO/action/author_attestation","sign_citation":"https://pith.science/pith/P2XFTJ5KDN5Z2BAF4J67FLLEHO/action/citation_signature","submit_replication":"https://pith.science/pith/P2XFTJ5KDN5Z2BAF4J67FLLEHO/action/replication_record"}},"created_at":"2026-07-05T10:57:23.677389+00:00","updated_at":"2026-07-05T10:57:23.677389+00:00"}