{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:APIXJFSO2ZHOPMTTMXPRL3USN7","short_pith_number":"pith:APIXJFSO","schema_version":"1.0","canonical_sha256":"03d174964ed64ee7b27365df15ee926fca4e64bb15efe0febdf22399f131c0fc","source":{"kind":"arxiv","id":"2606.18531","version":1},"attestation_state":"computed","paper":{"title":"When Does Trajectory-Level Supervision Permit Efficient Offline Reinforcement Learning?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Tengyang Xie, Xuanfei Ren","submitted_at":"2026-06-16T22:55:45Z","abstract_excerpt":"Offline reinforcement learning is typically analyzed under process-level reward supervision, yet many sequential decision datasets\n  record only trajectory-level outcomes. We develop a statistical theory for offline policy optimization from such outcome-level\n  supervision. We first study the canonical setting where the target remains the expected cumulative reward, but each offline trajectory\n  provides only a scalar label whose conditional mean is the cumulative return. We propose OPAC, a pessimistic actor-critic algorithm\n  that learns a latent reward model and optimizes a policy from traje"},"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":"2606.18531","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2026-06-16T22:55:45Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"12bc0919d720f8f0d32b4984806e1ffa90717be3e15e26f080782522f76e909f","abstract_canon_sha256":"992acb844f7852103d68a02a7cb2497aa81171f3bb399a2db6640c73ac361f20"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-19T16:11:40.218776Z","signature_b64":"z8JfU2zdDWdbtSquNuo8ZC7Wf+aiBWHDI/kl4GKdXw9Ov+GHNNKa4E5/9uL7tPxGyFBKHjsZekz6r/OTT9d1BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"03d174964ed64ee7b27365df15ee926fca4e64bb15efe0febdf22399f131c0fc","last_reissued_at":"2026-06-19T16:11:40.218431Z","signature_status":"signed_v1","first_computed_at":"2026-06-19T16:11:40.218431Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"When Does Trajectory-Level Supervision Permit Efficient Offline Reinforcement Learning?","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Tengyang Xie, Xuanfei Ren","submitted_at":"2026-06-16T22:55:45Z","abstract_excerpt":"Offline reinforcement learning is typically analyzed under process-level reward supervision, yet many sequential decision datasets\n  record only trajectory-level outcomes. We develop a statistical theory for offline policy optimization from such outcome-level\n  supervision. We first study the canonical setting where the target remains the expected cumulative reward, but each offline trajectory\n  provides only a scalar label whose conditional mean is the cumulative return. We propose OPAC, a pessimistic actor-critic algorithm\n  that learns a latent reward model and optimizes a policy from traje"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.18531","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/2606.18531/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":"2606.18531","created_at":"2026-06-19T16:11:40.218486+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.18531v1","created_at":"2026-06-19T16:11:40.218486+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.18531","created_at":"2026-06-19T16:11:40.218486+00:00"},{"alias_kind":"pith_short_12","alias_value":"APIXJFSO2ZHO","created_at":"2026-06-19T16:11:40.218486+00:00"},{"alias_kind":"pith_short_16","alias_value":"APIXJFSO2ZHOPMTT","created_at":"2026-06-19T16:11:40.218486+00:00"},{"alias_kind":"pith_short_8","alias_value":"APIXJFSO","created_at":"2026-06-19T16:11:40.218486+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/APIXJFSO2ZHOPMTTMXPRL3USN7","json":"https://pith.science/pith/APIXJFSO2ZHOPMTTMXPRL3USN7.json","graph_json":"https://pith.science/api/pith-number/APIXJFSO2ZHOPMTTMXPRL3USN7/graph.json","events_json":"https://pith.science/api/pith-number/APIXJFSO2ZHOPMTTMXPRL3USN7/events.json","paper":"https://pith.science/paper/APIXJFSO"},"agent_actions":{"view_html":"https://pith.science/pith/APIXJFSO2ZHOPMTTMXPRL3USN7","download_json":"https://pith.science/pith/APIXJFSO2ZHOPMTTMXPRL3USN7.json","view_paper":"https://pith.science/paper/APIXJFSO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.18531&json=true","fetch_graph":"https://pith.science/api/pith-number/APIXJFSO2ZHOPMTTMXPRL3USN7/graph.json","fetch_events":"https://pith.science/api/pith-number/APIXJFSO2ZHOPMTTMXPRL3USN7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/APIXJFSO2ZHOPMTTMXPRL3USN7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/APIXJFSO2ZHOPMTTMXPRL3USN7/action/storage_attestation","attest_author":"https://pith.science/pith/APIXJFSO2ZHOPMTTMXPRL3USN7/action/author_attestation","sign_citation":"https://pith.science/pith/APIXJFSO2ZHOPMTTMXPRL3USN7/action/citation_signature","submit_replication":"https://pith.science/pith/APIXJFSO2ZHOPMTTMXPRL3USN7/action/replication_record"}},"created_at":"2026-06-19T16:11:40.218486+00:00","updated_at":"2026-06-19T16:11:40.218486+00:00"}