{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:3QS5INCNS7RLZFGAMED5TQ7ZK2","short_pith_number":"pith:3QS5INCN","schema_version":"1.0","canonical_sha256":"dc25d4344d97e2bc94c06107d9c3f9569cffcb97c8a553186f41a2dbda094885","source":{"kind":"arxiv","id":"2210.00999","version":2},"attestation_state":"computed","paper":{"title":"Latent State Marginalization as a Low-cost Approach for Improving Exploration","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Aaron Courville, Amy Zhang, Dinghuai Zhang, Qinqing Zheng, Ricky T. Q. Chen, Yoshua Bengio","submitted_at":"2022-10-03T15:09:12Z","abstract_excerpt":"While the maximum entropy (MaxEnt) reinforcement learning (RL) framework -- often touted for its exploration and robustness capabilities -- is usually motivated from a probabilistic perspective, the use of deep probabilistic models has not gained much traction in practice due to their inherent complexity. In this work, we propose the adoption of latent variable policies within the MaxEnt framework, which we show can provably approximate any policy distribution, and additionally, naturally emerges under the use of world models with a latent belief state. We discuss why latent variable policies "},"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":"2210.00999","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2022-10-03T15:09:12Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"e9ffbad9f90915d07989989468f5d2ff7071569a6f5661babea0a15a1971ef7b","abstract_canon_sha256":"9da1e1723485904788bb0041aeb4db7b122f9db164ffdb22f8a4f22385dd2460"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:40:32.449262Z","signature_b64":"4e1oLSNiMqo8OMdgisKVHEs2xSttaf447ChHwJjC+zilmI6hoqj1iSjuiwvL5Vnctmef5ErTHrNZEjC2091MCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dc25d4344d97e2bc94c06107d9c3f9569cffcb97c8a553186f41a2dbda094885","last_reissued_at":"2026-07-05T05:40:32.448809Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:40:32.448809Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Latent State Marginalization as a Low-cost Approach for Improving Exploration","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Aaron Courville, Amy Zhang, Dinghuai Zhang, Qinqing Zheng, Ricky T. Q. Chen, Yoshua Bengio","submitted_at":"2022-10-03T15:09:12Z","abstract_excerpt":"While the maximum entropy (MaxEnt) reinforcement learning (RL) framework -- often touted for its exploration and robustness capabilities -- is usually motivated from a probabilistic perspective, the use of deep probabilistic models has not gained much traction in practice due to their inherent complexity. In this work, we propose the adoption of latent variable policies within the MaxEnt framework, which we show can provably approximate any policy distribution, and additionally, naturally emerges under the use of world models with a latent belief state. We discuss why latent variable policies "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2210.00999","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/2210.00999/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":"2210.00999","created_at":"2026-07-05T05:40:32.448865+00:00"},{"alias_kind":"arxiv_version","alias_value":"2210.00999v2","created_at":"2026-07-05T05:40:32.448865+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2210.00999","created_at":"2026-07-05T05:40:32.448865+00:00"},{"alias_kind":"pith_short_12","alias_value":"3QS5INCNS7RL","created_at":"2026-07-05T05:40:32.448865+00:00"},{"alias_kind":"pith_short_16","alias_value":"3QS5INCNS7RLZFGA","created_at":"2026-07-05T05:40:32.448865+00:00"},{"alias_kind":"pith_short_8","alias_value":"3QS5INCN","created_at":"2026-07-05T05:40:32.448865+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.16473","citing_title":"Learning Temporal Abstractions via Variational Homomorphisms in Option-Induced Abstract MDPs","ref_index":68,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3QS5INCNS7RLZFGAMED5TQ7ZK2","json":"https://pith.science/pith/3QS5INCNS7RLZFGAMED5TQ7ZK2.json","graph_json":"https://pith.science/api/pith-number/3QS5INCNS7RLZFGAMED5TQ7ZK2/graph.json","events_json":"https://pith.science/api/pith-number/3QS5INCNS7RLZFGAMED5TQ7ZK2/events.json","paper":"https://pith.science/paper/3QS5INCN"},"agent_actions":{"view_html":"https://pith.science/pith/3QS5INCNS7RLZFGAMED5TQ7ZK2","download_json":"https://pith.science/pith/3QS5INCNS7RLZFGAMED5TQ7ZK2.json","view_paper":"https://pith.science/paper/3QS5INCN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2210.00999&json=true","fetch_graph":"https://pith.science/api/pith-number/3QS5INCNS7RLZFGAMED5TQ7ZK2/graph.json","fetch_events":"https://pith.science/api/pith-number/3QS5INCNS7RLZFGAMED5TQ7ZK2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3QS5INCNS7RLZFGAMED5TQ7ZK2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3QS5INCNS7RLZFGAMED5TQ7ZK2/action/storage_attestation","attest_author":"https://pith.science/pith/3QS5INCNS7RLZFGAMED5TQ7ZK2/action/author_attestation","sign_citation":"https://pith.science/pith/3QS5INCNS7RLZFGAMED5TQ7ZK2/action/citation_signature","submit_replication":"https://pith.science/pith/3QS5INCNS7RLZFGAMED5TQ7ZK2/action/replication_record"}},"created_at":"2026-07-05T05:40:32.448865+00:00","updated_at":"2026-07-05T05:40:32.448865+00:00"}