{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:LR5NN2W2SU7O57LJT5HJLIMXV3","short_pith_number":"pith:LR5NN2W2","schema_version":"1.0","canonical_sha256":"5c7ad6eada953eeefd699f4e95a197aed9563fdd41901618d0611db28c795d41","source":{"kind":"arxiv","id":"2305.17083","version":2},"attestation_state":"computed","paper":{"title":"A Policy Gradient Method for Confounded POMDPs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","econ.EM","math.ST","stat.ME","stat.TH"],"primary_cat":"stat.ML","authors_text":"Mao Hong, Yanxun Xu, Zhengling Qi","submitted_at":"2023-05-26T16:48:05Z","abstract_excerpt":"In this paper, we propose a policy gradient method for confounded partially observable Markov decision processes (POMDPs) with continuous state and observation spaces in the offline setting. We first establish a novel identification result to non-parametrically estimate any history-dependent policy gradient under POMDPs using the offline data. The identification enables us to solve a sequence of conditional moment restrictions and adopt the min-max learning procedure with general function approximation for estimating the policy gradient. We then provide a finite-sample non-asymptotic bound for"},"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":"2305.17083","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2023-05-26T16:48:05Z","cross_cats_sorted":["cs.LG","econ.EM","math.ST","stat.ME","stat.TH"],"title_canon_sha256":"92866e1bfd42b2f93406467122a09b40da7b9fad33a4dda4b0f7eb5ac5a617ae","abstract_canon_sha256":"d84a06ae0777deb7c93beeea3cf542aebbf55b4d72d2a5f83d90ddc359cbd238"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:18:58.050699Z","signature_b64":"rUns8wAXa6NbUSi1olqsVqYq8fqz1AoCleC3OqvhGRK4qC64aC4DHljY1DlbgD8/fr7HxyEhJ8sPS4rUI7bCAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5c7ad6eada953eeefd699f4e95a197aed9563fdd41901618d0611db28c795d41","last_reissued_at":"2026-07-05T07:18:58.050176Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:18:58.050176Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Policy Gradient Method for Confounded POMDPs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","econ.EM","math.ST","stat.ME","stat.TH"],"primary_cat":"stat.ML","authors_text":"Mao Hong, Yanxun Xu, Zhengling Qi","submitted_at":"2023-05-26T16:48:05Z","abstract_excerpt":"In this paper, we propose a policy gradient method for confounded partially observable Markov decision processes (POMDPs) with continuous state and observation spaces in the offline setting. We first establish a novel identification result to non-parametrically estimate any history-dependent policy gradient under POMDPs using the offline data. The identification enables us to solve a sequence of conditional moment restrictions and adopt the min-max learning procedure with general function approximation for estimating the policy gradient. We then provide a finite-sample non-asymptotic bound for"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.17083","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/2305.17083/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":"2305.17083","created_at":"2026-07-05T07:18:58.050229+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.17083v2","created_at":"2026-07-05T07:18:58.050229+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.17083","created_at":"2026-07-05T07:18:58.050229+00:00"},{"alias_kind":"pith_short_12","alias_value":"LR5NN2W2SU7O","created_at":"2026-07-05T07:18:58.050229+00:00"},{"alias_kind":"pith_short_16","alias_value":"LR5NN2W2SU7O57LJ","created_at":"2026-07-05T07:18:58.050229+00:00"},{"alias_kind":"pith_short_8","alias_value":"LR5NN2W2","created_at":"2026-07-05T07:18:58.050229+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.20206","citing_title":"Off-Policy Evaluation for Missingness-Aware Policies in MDPs with Rewards Missing Not at Random","ref_index":17,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LR5NN2W2SU7O57LJT5HJLIMXV3","json":"https://pith.science/pith/LR5NN2W2SU7O57LJT5HJLIMXV3.json","graph_json":"https://pith.science/api/pith-number/LR5NN2W2SU7O57LJT5HJLIMXV3/graph.json","events_json":"https://pith.science/api/pith-number/LR5NN2W2SU7O57LJT5HJLIMXV3/events.json","paper":"https://pith.science/paper/LR5NN2W2"},"agent_actions":{"view_html":"https://pith.science/pith/LR5NN2W2SU7O57LJT5HJLIMXV3","download_json":"https://pith.science/pith/LR5NN2W2SU7O57LJT5HJLIMXV3.json","view_paper":"https://pith.science/paper/LR5NN2W2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.17083&json=true","fetch_graph":"https://pith.science/api/pith-number/LR5NN2W2SU7O57LJT5HJLIMXV3/graph.json","fetch_events":"https://pith.science/api/pith-number/LR5NN2W2SU7O57LJT5HJLIMXV3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LR5NN2W2SU7O57LJT5HJLIMXV3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LR5NN2W2SU7O57LJT5HJLIMXV3/action/storage_attestation","attest_author":"https://pith.science/pith/LR5NN2W2SU7O57LJT5HJLIMXV3/action/author_attestation","sign_citation":"https://pith.science/pith/LR5NN2W2SU7O57LJT5HJLIMXV3/action/citation_signature","submit_replication":"https://pith.science/pith/LR5NN2W2SU7O57LJT5HJLIMXV3/action/replication_record"}},"created_at":"2026-07-05T07:18:58.050229+00:00","updated_at":"2026-07-05T07:18:58.050229+00:00"}