{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:D5JZDU3KI6GZDTWMR76IAIL54I","short_pith_number":"pith:D5JZDU3K","schema_version":"1.0","canonical_sha256":"1f5391d36a478d91cecc8ffc80217de20c04d3e068e84f35c82addad6e202746","source":{"kind":"arxiv","id":"2605.27877","version":1},"attestation_state":"computed","paper":{"title":"SPAR: Support-Preserving Action Rectification","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Binbin Lin, Jiaxin Zhao, Weihang Pan, Xun Liang","submitted_at":"2026-05-27T02:53:41Z","abstract_excerpt":"Offline policy improvement faces an inherent conflict between maximizing value and fitting the data distribution. While in-sample weighted regression is stable, it suffers from over-conservatism that suppresses high-value actions in the distribution tail; conversely, gradient-based approaches often exhibit a fitting-optimization conflict of gradients, which drives the policy off the data manifold. To address this, we propose Support-Preserving Action Rectification (SPAR), which reframes global learning as a local residual rectification anchored to a frozen pure behavior cloning policy. This fr"},"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":"2605.27877","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-27T02:53:41Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"014c2a6f4bdb19efc435f3a95380fa018ca96cbd2fec458f7be72515ba6d98e2","abstract_canon_sha256":"cd0f367cb9e0e419abc8e6fe47437846251af0172b068ed7915af84cab1f2425"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T01:04:51.076264Z","signature_b64":"olQl2quC/zSHdtTuBf0n458bmM5+tOAgIWaEVje0RQ7Cp6wrJSOh/XZ+Gpe2AsEZvxVU8Z+dIYWDWf9Fby2yBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1f5391d36a478d91cecc8ffc80217de20c04d3e068e84f35c82addad6e202746","last_reissued_at":"2026-05-28T01:04:51.075873Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T01:04:51.075873Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SPAR: Support-Preserving Action Rectification","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Binbin Lin, Jiaxin Zhao, Weihang Pan, Xun Liang","submitted_at":"2026-05-27T02:53:41Z","abstract_excerpt":"Offline policy improvement faces an inherent conflict between maximizing value and fitting the data distribution. While in-sample weighted regression is stable, it suffers from over-conservatism that suppresses high-value actions in the distribution tail; conversely, gradient-based approaches often exhibit a fitting-optimization conflict of gradients, which drives the policy off the data manifold. To address this, we propose Support-Preserving Action Rectification (SPAR), which reframes global learning as a local residual rectification anchored to a frozen pure behavior cloning policy. This fr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.27877","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/2605.27877/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":"2605.27877","created_at":"2026-05-28T01:04:51.075943+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.27877v1","created_at":"2026-05-28T01:04:51.075943+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.27877","created_at":"2026-05-28T01:04:51.075943+00:00"},{"alias_kind":"pith_short_12","alias_value":"D5JZDU3KI6GZ","created_at":"2026-05-28T01:04:51.075943+00:00"},{"alias_kind":"pith_short_16","alias_value":"D5JZDU3KI6GZDTWM","created_at":"2026-05-28T01:04:51.075943+00:00"},{"alias_kind":"pith_short_8","alias_value":"D5JZDU3K","created_at":"2026-05-28T01:04:51.075943+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/D5JZDU3KI6GZDTWMR76IAIL54I","json":"https://pith.science/pith/D5JZDU3KI6GZDTWMR76IAIL54I.json","graph_json":"https://pith.science/api/pith-number/D5JZDU3KI6GZDTWMR76IAIL54I/graph.json","events_json":"https://pith.science/api/pith-number/D5JZDU3KI6GZDTWMR76IAIL54I/events.json","paper":"https://pith.science/paper/D5JZDU3K"},"agent_actions":{"view_html":"https://pith.science/pith/D5JZDU3KI6GZDTWMR76IAIL54I","download_json":"https://pith.science/pith/D5JZDU3KI6GZDTWMR76IAIL54I.json","view_paper":"https://pith.science/paper/D5JZDU3K","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.27877&json=true","fetch_graph":"https://pith.science/api/pith-number/D5JZDU3KI6GZDTWMR76IAIL54I/graph.json","fetch_events":"https://pith.science/api/pith-number/D5JZDU3KI6GZDTWMR76IAIL54I/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/D5JZDU3KI6GZDTWMR76IAIL54I/action/timestamp_anchor","attest_storage":"https://pith.science/pith/D5JZDU3KI6GZDTWMR76IAIL54I/action/storage_attestation","attest_author":"https://pith.science/pith/D5JZDU3KI6GZDTWMR76IAIL54I/action/author_attestation","sign_citation":"https://pith.science/pith/D5JZDU3KI6GZDTWMR76IAIL54I/action/citation_signature","submit_replication":"https://pith.science/pith/D5JZDU3KI6GZDTWMR76IAIL54I/action/replication_record"}},"created_at":"2026-05-28T01:04:51.075943+00:00","updated_at":"2026-05-28T01:04:51.075943+00:00"}