{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:7VN7NXT5Z7DH45VYIAYP6EYLQS","short_pith_number":"pith:7VN7NXT5","schema_version":"1.0","canonical_sha256":"fd5bf6de7dcfc67e76b84030ff130b8484c48b6433b03c5d7b048b4fd77d7d58","source":{"kind":"arxiv","id":"2406.19861","version":2},"attestation_state":"computed","paper":{"title":"Operator World Models for Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Carlo Ciliberto, Marco Prattic\\`o, Massimiliano Pontil, Pietro Novelli","submitted_at":"2024-06-28T12:05:47Z","abstract_excerpt":"Policy Mirror Descent (PMD) is a powerful and theoretically sound methodology for sequential decision-making. However, it is not directly applicable to Reinforcement Learning (RL) due to the inaccessibility of explicit action-value functions. We address this challenge by introducing a novel approach based on learning a world model of the environment using conditional mean embeddings. Leveraging tools from operator theory we derive a closed-form expression of the action-value function in terms of the world model via simple matrix operations. Combining these estimators with PMD leads to POWR, a "},"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":"2406.19861","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-06-28T12:05:47Z","cross_cats_sorted":["math.OC","stat.ML"],"title_canon_sha256":"9066639295aaa54aeb76cd880aaf0beba87fb481a11d127dbc4b21103511fd76","abstract_canon_sha256":"b9091d9fba3798ce570eb7254c9ab5bc2ebb13faba528838a1c72234f90fdec5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:28:52.534683Z","signature_b64":"+Pw+xlImRRmsTcV60U6K6a7jdJDoSQA9cpbkJsF07t/QnHN3Y83/bj/mV3Rj3A8mVvoduznmwYMnTyJz2QkcBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fd5bf6de7dcfc67e76b84030ff130b8484c48b6433b03c5d7b048b4fd77d7d58","last_reissued_at":"2026-07-05T09:28:52.534167Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:28:52.534167Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Operator World Models for Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Carlo Ciliberto, Marco Prattic\\`o, Massimiliano Pontil, Pietro Novelli","submitted_at":"2024-06-28T12:05:47Z","abstract_excerpt":"Policy Mirror Descent (PMD) is a powerful and theoretically sound methodology for sequential decision-making. However, it is not directly applicable to Reinforcement Learning (RL) due to the inaccessibility of explicit action-value functions. We address this challenge by introducing a novel approach based on learning a world model of the environment using conditional mean embeddings. Leveraging tools from operator theory we derive a closed-form expression of the action-value function in terms of the world model via simple matrix operations. Combining these estimators with PMD leads to POWR, a "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.19861","kind":"arxiv","version":2},"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/2406.19861/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":"2406.19861","created_at":"2026-07-05T09:28:52.534223+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.19861v2","created_at":"2026-07-05T09:28:52.534223+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.19861","created_at":"2026-07-05T09:28:52.534223+00:00"},{"alias_kind":"pith_short_12","alias_value":"7VN7NXT5Z7DH","created_at":"2026-07-05T09:28:52.534223+00:00"},{"alias_kind":"pith_short_16","alias_value":"7VN7NXT5Z7DH45VY","created_at":"2026-07-05T09:28:52.534223+00:00"},{"alias_kind":"pith_short_8","alias_value":"7VN7NXT5","created_at":"2026-07-05T09:28:52.534223+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.18671","citing_title":"Self-Supervised Evolution Operator Learning for High-Dimensional Dynamical Systems","ref_index":48,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7VN7NXT5Z7DH45VYIAYP6EYLQS","json":"https://pith.science/pith/7VN7NXT5Z7DH45VYIAYP6EYLQS.json","graph_json":"https://pith.science/api/pith-number/7VN7NXT5Z7DH45VYIAYP6EYLQS/graph.json","events_json":"https://pith.science/api/pith-number/7VN7NXT5Z7DH45VYIAYP6EYLQS/events.json","paper":"https://pith.science/paper/7VN7NXT5"},"agent_actions":{"view_html":"https://pith.science/pith/7VN7NXT5Z7DH45VYIAYP6EYLQS","download_json":"https://pith.science/pith/7VN7NXT5Z7DH45VYIAYP6EYLQS.json","view_paper":"https://pith.science/paper/7VN7NXT5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.19861&json=true","fetch_graph":"https://pith.science/api/pith-number/7VN7NXT5Z7DH45VYIAYP6EYLQS/graph.json","fetch_events":"https://pith.science/api/pith-number/7VN7NXT5Z7DH45VYIAYP6EYLQS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7VN7NXT5Z7DH45VYIAYP6EYLQS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7VN7NXT5Z7DH45VYIAYP6EYLQS/action/storage_attestation","attest_author":"https://pith.science/pith/7VN7NXT5Z7DH45VYIAYP6EYLQS/action/author_attestation","sign_citation":"https://pith.science/pith/7VN7NXT5Z7DH45VYIAYP6EYLQS/action/citation_signature","submit_replication":"https://pith.science/pith/7VN7NXT5Z7DH45VYIAYP6EYLQS/action/replication_record"}},"created_at":"2026-07-05T09:28:52.534223+00:00","updated_at":"2026-07-05T09:28:52.534223+00:00"}