{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:V7WRVDCJFOR2VKGYBB5QUJBQ5F","short_pith_number":"pith:V7WRVDCJ","schema_version":"1.0","canonical_sha256":"afed1a8c492ba3aaa8d8087b0a2430e962a524fbe8bac81a91897fb296e68089","source":{"kind":"arxiv","id":"2608.09233","version":1},"attestation_state":"computed","paper":{"title":"DreOPD: Degraded-Reference Extrapolative On-Policy Distillation for Flow-matching Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Chengfei Cai, Liang Han, Lin Xu, Mingfeng Lin, Yuxiang Wei","submitted_at":"2026-08-10T07:55:57Z","abstract_excerpt":"Flow-matching models are now a mainstream method to image generation, but its adaptation to diverse downstream scenarios typically relies on post-training, which may cause conflicts among task-specific optimization objectives. Reinforcement learning enables direct optimization of task-specific rewards beyond the original models, yet trajectory-level optimization may incur high-variance gradients and cross-task interference. On-policy distillation (OPD) offers dense and stable supervision on student rollouts, but conventional teacher matching remains imitation-based. We propose DreOPD, a Degrad"},"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":"2608.09233","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-08-10T07:55:57Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"286469375bd170b7ac6fdb6ee8bed9ecec67082ccc904f2a54f958d73689048a","abstract_canon_sha256":"e39ad5a1bd3c64dd60ccd3861776dc446f550ce6ee2f0b04ab0f53af60651af4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-11T02:22:08.549046Z","signature_b64":"8aKVKa97plfJQW5NMy0FguRzhoYslapOICUAnV8rgGpspm2xGCCqpN+nLwN7ZCDoLyL42ikchdKltwgCaiPLBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"afed1a8c492ba3aaa8d8087b0a2430e962a524fbe8bac81a91897fb296e68089","last_reissued_at":"2026-08-11T02:22:08.547321Z","signature_status":"signed_v1","first_computed_at":"2026-08-11T02:22:08.547321Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DreOPD: Degraded-Reference Extrapolative On-Policy Distillation for Flow-matching Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Chengfei Cai, Liang Han, Lin Xu, Mingfeng Lin, Yuxiang Wei","submitted_at":"2026-08-10T07:55:57Z","abstract_excerpt":"Flow-matching models are now a mainstream method to image generation, but its adaptation to diverse downstream scenarios typically relies on post-training, which may cause conflicts among task-specific optimization objectives. Reinforcement learning enables direct optimization of task-specific rewards beyond the original models, yet trajectory-level optimization may incur high-variance gradients and cross-task interference. On-policy distillation (OPD) offers dense and stable supervision on student rollouts, but conventional teacher matching remains imitation-based. We propose DreOPD, a Degrad"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.09233","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/2608.09233/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":"2608.09233","created_at":"2026-08-11T02:22:08.548052+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.09233v1","created_at":"2026-08-11T02:22:08.548052+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.09233","created_at":"2026-08-11T02:22:08.548052+00:00"},{"alias_kind":"pith_short_12","alias_value":"V7WRVDCJFOR2","created_at":"2026-08-11T02:22:08.548052+00:00"},{"alias_kind":"pith_short_16","alias_value":"V7WRVDCJFOR2VKGY","created_at":"2026-08-11T02:22:08.548052+00:00"},{"alias_kind":"pith_short_8","alias_value":"V7WRVDCJ","created_at":"2026-08-11T02:22:08.548052+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/V7WRVDCJFOR2VKGYBB5QUJBQ5F","json":"https://pith.science/pith/V7WRVDCJFOR2VKGYBB5QUJBQ5F.json","graph_json":"https://pith.science/api/pith-number/V7WRVDCJFOR2VKGYBB5QUJBQ5F/graph.json","events_json":"https://pith.science/api/pith-number/V7WRVDCJFOR2VKGYBB5QUJBQ5F/events.json","paper":"https://pith.science/paper/V7WRVDCJ"},"agent_actions":{"view_html":"https://pith.science/pith/V7WRVDCJFOR2VKGYBB5QUJBQ5F","download_json":"https://pith.science/pith/V7WRVDCJFOR2VKGYBB5QUJBQ5F.json","view_paper":"https://pith.science/paper/V7WRVDCJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.09233&json=true","fetch_graph":"https://pith.science/api/pith-number/V7WRVDCJFOR2VKGYBB5QUJBQ5F/graph.json","fetch_events":"https://pith.science/api/pith-number/V7WRVDCJFOR2VKGYBB5QUJBQ5F/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/V7WRVDCJFOR2VKGYBB5QUJBQ5F/action/timestamp_anchor","attest_storage":"https://pith.science/pith/V7WRVDCJFOR2VKGYBB5QUJBQ5F/action/storage_attestation","attest_author":"https://pith.science/pith/V7WRVDCJFOR2VKGYBB5QUJBQ5F/action/author_attestation","sign_citation":"https://pith.science/pith/V7WRVDCJFOR2VKGYBB5QUJBQ5F/action/citation_signature","submit_replication":"https://pith.science/pith/V7WRVDCJFOR2VKGYBB5QUJBQ5F/action/replication_record"}},"created_at":"2026-08-11T02:22:08.548052+00:00","updated_at":"2026-08-11T02:22:08.548052+00:00"}