{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:TRLBGPV3KK2XFA6DXFOYEDJJTJ","short_pith_number":"pith:TRLBGPV3","schema_version":"1.0","canonical_sha256":"9c56133ebb52b57283c3b95d820d299a648b416e55f22bb610d060aa4cddce4b","source":{"kind":"arxiv","id":"2110.05038","version":3},"attestation_state":"computed","paper":{"title":"Recurrent Model-Free RL Can Be a Strong Baseline for Many POMDPs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.RO"],"primary_cat":"cs.LG","authors_text":"Benjamin Eysenbach, Ruslan Salakhutdinov, Tianwei Ni","submitted_at":"2021-10-11T07:09:14Z","abstract_excerpt":"Many problems in RL, such as meta-RL, robust RL, generalization in RL, and temporal credit assignment, can be cast as POMDPs. In theory, simply augmenting model-free RL with memory-based architectures, such as recurrent neural networks, provides a general approach to solving all types of POMDPs. However, prior work has found that such recurrent model-free RL methods tend to perform worse than more specialized algorithms that are designed for specific types of POMDPs. This paper revisits this claim. We find that careful architecture and hyperparameter decisions can often yield a recurrent model"},"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":"2110.05038","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-10-11T07:09:14Z","cross_cats_sorted":["cs.AI","cs.RO"],"title_canon_sha256":"ad606edaa6cee0b713ba2e051e4c8457dfe47623089f6fad383b791008c18bd3","abstract_canon_sha256":"bc91bf895cd78a4c1fae0cae70c03eb070540a1640a7ab3e2db59a46ae3256f1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:28:52.994837Z","signature_b64":"mnuYgLc+4ArrNBdjnzOqR5iKmcV5vSjdh1XF+z1hfnZbVe2AtnafTPwE9ZR7zR3kG7cwKwwywFPWuefauF2jBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9c56133ebb52b57283c3b95d820d299a648b416e55f22bb610d060aa4cddce4b","last_reissued_at":"2026-07-05T04:28:52.994359Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:28:52.994359Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Recurrent Model-Free RL Can Be a Strong Baseline for Many POMDPs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.RO"],"primary_cat":"cs.LG","authors_text":"Benjamin Eysenbach, Ruslan Salakhutdinov, Tianwei Ni","submitted_at":"2021-10-11T07:09:14Z","abstract_excerpt":"Many problems in RL, such as meta-RL, robust RL, generalization in RL, and temporal credit assignment, can be cast as POMDPs. In theory, simply augmenting model-free RL with memory-based architectures, such as recurrent neural networks, provides a general approach to solving all types of POMDPs. However, prior work has found that such recurrent model-free RL methods tend to perform worse than more specialized algorithms that are designed for specific types of POMDPs. This paper revisits this claim. We find that careful architecture and hyperparameter decisions can often yield a recurrent model"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2110.05038","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/2110.05038/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":"2110.05038","created_at":"2026-07-05T04:28:52.994412+00:00"},{"alias_kind":"arxiv_version","alias_value":"2110.05038v3","created_at":"2026-07-05T04:28:52.994412+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2110.05038","created_at":"2026-07-05T04:28:52.994412+00:00"},{"alias_kind":"pith_short_12","alias_value":"TRLBGPV3KK2X","created_at":"2026-07-05T04:28:52.994412+00:00"},{"alias_kind":"pith_short_16","alias_value":"TRLBGPV3KK2XFA6D","created_at":"2026-07-05T04:28:52.994412+00:00"},{"alias_kind":"pith_short_8","alias_value":"TRLBGPV3","created_at":"2026-07-05T04:28:52.994412+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.18820","citing_title":"Maturing Markov Decision Processes: Decision Making under Increasing Information and Shrinking Action Sets","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2602.05048","citing_title":"MINT: Minimal Information Neuro-Symbolic Tree for Objective-Driven Knowledge-Gap Reasoning and Active Elicitation","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2602.19837","citing_title":"Meta-Learning and Meta-Reinforcement Learning -- Tracing the Path towards DeepMind's Adaptive Agent","ref_index":110,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13740","citing_title":"Learning POMDP World Models from Observations with Language-Model Priors","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2604.09671","citing_title":"Belief-State RWKV for Reinforcement Learning under Partial Observability","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02552","citing_title":"Recurrent Deep Reinforcement Learning for Chemotherapy Control under Partial Observability","ref_index":20,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TRLBGPV3KK2XFA6DXFOYEDJJTJ","json":"https://pith.science/pith/TRLBGPV3KK2XFA6DXFOYEDJJTJ.json","graph_json":"https://pith.science/api/pith-number/TRLBGPV3KK2XFA6DXFOYEDJJTJ/graph.json","events_json":"https://pith.science/api/pith-number/TRLBGPV3KK2XFA6DXFOYEDJJTJ/events.json","paper":"https://pith.science/paper/TRLBGPV3"},"agent_actions":{"view_html":"https://pith.science/pith/TRLBGPV3KK2XFA6DXFOYEDJJTJ","download_json":"https://pith.science/pith/TRLBGPV3KK2XFA6DXFOYEDJJTJ.json","view_paper":"https://pith.science/paper/TRLBGPV3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2110.05038&json=true","fetch_graph":"https://pith.science/api/pith-number/TRLBGPV3KK2XFA6DXFOYEDJJTJ/graph.json","fetch_events":"https://pith.science/api/pith-number/TRLBGPV3KK2XFA6DXFOYEDJJTJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TRLBGPV3KK2XFA6DXFOYEDJJTJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TRLBGPV3KK2XFA6DXFOYEDJJTJ/action/storage_attestation","attest_author":"https://pith.science/pith/TRLBGPV3KK2XFA6DXFOYEDJJTJ/action/author_attestation","sign_citation":"https://pith.science/pith/TRLBGPV3KK2XFA6DXFOYEDJJTJ/action/citation_signature","submit_replication":"https://pith.science/pith/TRLBGPV3KK2XFA6DXFOYEDJJTJ/action/replication_record"}},"created_at":"2026-07-05T04:28:52.994412+00:00","updated_at":"2026-07-05T04:28:52.994412+00:00"}