{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:ZTXKHUTEGVZXRJFIBR75J32VZ3","short_pith_number":"pith:ZTXKHUTE","schema_version":"1.0","canonical_sha256":"cceea3d264357378a4a80c7fd4ef55ced59f4967ad1169f3700246b0bf896cf1","source":{"kind":"arxiv","id":"2605.24001","version":1},"attestation_state":"computed","paper":{"title":"Diff-Instruct with Diffused Reward: Towards Principled One-step Generator RL","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Guang Lin Haoyang Zheng Runzhe Zhang Guang Lin, Haoyang Zheng, Junyi Wu, Runzhe Zhang, Weijian Luo","submitted_at":"2026-05-18T21:26:33Z","abstract_excerpt":"Recent advances in one-step text-to-image generation have enabled real-time synthesis with remarkable efficiency and quality. Previous reinforcement learning methods for one-step generators combine image-space reward optimization with diffusion noisy-space distribution matching. This paradigm brings challenges due to a mismatch between terminal reward optimization and the underlying generative dynamics. As a result, optimization tends to exploit stochastic degrees of freedom, often improving reward at the expense of image fidelity. To address this issue, we propose Diff-Instruct with Diffused "},"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":"2605.24001","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-18T21:26:33Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"8b693e3e46c18478e846511ee445dc892707634895c949ee40c2df94791f1255","abstract_canon_sha256":"bf1ae82645a0c8d5742462f7afedc9e417cb7ac2347fec1f64e26a0a1310f497"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-26T01:02:40.249564Z","signature_b64":"/b24lJhwISohzuvqtuYv7vkOnZ64PjGkH1h3Tf3cUAPlXYEevzb7b6/pxNfJ+3ys9e7NYSnvhoP8pqjALGwzBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cceea3d264357378a4a80c7fd4ef55ced59f4967ad1169f3700246b0bf896cf1","last_reissued_at":"2026-05-26T01:02:40.248884Z","signature_status":"signed_v1","first_computed_at":"2026-05-26T01:02:40.248884Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Diff-Instruct with Diffused Reward: Towards Principled One-step Generator RL","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Guang Lin Haoyang Zheng Runzhe Zhang Guang Lin, Haoyang Zheng, Junyi Wu, Runzhe Zhang, Weijian Luo","submitted_at":"2026-05-18T21:26:33Z","abstract_excerpt":"Recent advances in one-step text-to-image generation have enabled real-time synthesis with remarkable efficiency and quality. Previous reinforcement learning methods for one-step generators combine image-space reward optimization with diffusion noisy-space distribution matching. This paradigm brings challenges due to a mismatch between terminal reward optimization and the underlying generative dynamics. As a result, optimization tends to exploit stochastic degrees of freedom, often improving reward at the expense of image fidelity. To address this issue, we propose Diff-Instruct with Diffused "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.24001","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/2605.24001/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":"2605.24001","created_at":"2026-05-26T01:02:40.248994+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.24001v1","created_at":"2026-05-26T01:02:40.248994+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.24001","created_at":"2026-05-26T01:02:40.248994+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZTXKHUTEGVZX","created_at":"2026-05-26T01:02:40.248994+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZTXKHUTEGVZXRJFI","created_at":"2026-05-26T01:02:40.248994+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZTXKHUTE","created_at":"2026-05-26T01:02:40.248994+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/ZTXKHUTEGVZXRJFIBR75J32VZ3","json":"https://pith.science/pith/ZTXKHUTEGVZXRJFIBR75J32VZ3.json","graph_json":"https://pith.science/api/pith-number/ZTXKHUTEGVZXRJFIBR75J32VZ3/graph.json","events_json":"https://pith.science/api/pith-number/ZTXKHUTEGVZXRJFIBR75J32VZ3/events.json","paper":"https://pith.science/paper/ZTXKHUTE"},"agent_actions":{"view_html":"https://pith.science/pith/ZTXKHUTEGVZXRJFIBR75J32VZ3","download_json":"https://pith.science/pith/ZTXKHUTEGVZXRJFIBR75J32VZ3.json","view_paper":"https://pith.science/paper/ZTXKHUTE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.24001&json=true","fetch_graph":"https://pith.science/api/pith-number/ZTXKHUTEGVZXRJFIBR75J32VZ3/graph.json","fetch_events":"https://pith.science/api/pith-number/ZTXKHUTEGVZXRJFIBR75J32VZ3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZTXKHUTEGVZXRJFIBR75J32VZ3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZTXKHUTEGVZXRJFIBR75J32VZ3/action/storage_attestation","attest_author":"https://pith.science/pith/ZTXKHUTEGVZXRJFIBR75J32VZ3/action/author_attestation","sign_citation":"https://pith.science/pith/ZTXKHUTEGVZXRJFIBR75J32VZ3/action/citation_signature","submit_replication":"https://pith.science/pith/ZTXKHUTEGVZXRJFIBR75J32VZ3/action/replication_record"}},"created_at":"2026-05-26T01:02:40.248994+00:00","updated_at":"2026-05-26T01:02:40.248994+00:00"}