{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:IOVTADSAIQNRQIZOM5ICML5RTE","short_pith_number":"pith:IOVTADSA","schema_version":"1.0","canonical_sha256":"43ab300e40441b18232e6750262fb19903d11d9b06c5390de10b5769fee83e8f","source":{"kind":"arxiv","id":"2604.17415","version":3},"attestation_state":"computed","paper":{"title":"Reward Score Matching: Unifying Reward-based Fine-tuning for Flow and Diffusion Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Many reward-based fine-tuning methods for diffusion and flow models reduce to a single score-matching objective against a value-guided target.","cross_cats":["cs.AI","cs.CV"],"primary_cat":"cs.LG","authors_text":"Jeongjae Lee, Jeongsol Kim, Jinho Chang, Jong Chul Ye","submitted_at":"2026-04-19T12:47:52Z","abstract_excerpt":"Reward-based fine-tuning steers a pretrained diffusion or flow-based generative model toward higher-reward samples while remaining close to the pretrained model. Although existing methods are derived from different perspectives, we show that many can be written under a common framework, which we call reward score matching (RSM). Under this view, alignment becomes score matching against a value-guided target, and the main differences across methods reduce to the construction of the value-guidance estimator and the effective optimization strength across timesteps. This unification clarifies the "},"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":"2604.17415","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-04-19T12:47:52Z","cross_cats_sorted":["cs.AI","cs.CV"],"title_canon_sha256":"a9b9c9999d2b319856cd98f141b6c5316e3159c1ff90fcf1fc01e0069613e15f","abstract_canon_sha256":"965112703ecf90a21ea7fa74bb4649f763107130cbece4ac12a9ddcc00b3251d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T01:03:47.240129Z","signature_b64":"stWrP+g57acQKZrSwwom9zJ9sJUEl99Du3ZocIUXw5N3CFk70zaXftmv206rNfz+7dULLWpTV2bfZgmKU2VGAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"43ab300e40441b18232e6750262fb19903d11d9b06c5390de10b5769fee83e8f","last_reissued_at":"2026-06-02T01:03:47.239586Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T01:03:47.239586Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Reward Score Matching: Unifying Reward-based Fine-tuning for Flow and Diffusion Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Many reward-based fine-tuning methods for diffusion and flow models reduce to a single score-matching objective against a value-guided target.","cross_cats":["cs.AI","cs.CV"],"primary_cat":"cs.LG","authors_text":"Jeongjae Lee, Jeongsol Kim, Jinho Chang, Jong Chul Ye","submitted_at":"2026-04-19T12:47:52Z","abstract_excerpt":"Reward-based fine-tuning steers a pretrained diffusion or flow-based generative model toward higher-reward samples while remaining close to the pretrained model. Although existing methods are derived from different perspectives, we show that many can be written under a common framework, which we call reward score matching (RSM). Under this view, alignment becomes score matching against a value-guided target, and the main differences across methods reduce to the construction of the value-guidance estimator and the effective optimization strength across timesteps. This unification clarifies the "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Although existing methods are derived from different perspectives, we show that many can be written under a common framework, which we call reward score matching (RSM).","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the primary distinctions among existing methods reduce to the construction of the value-guidance estimator and the effective optimization strength across timesteps, without material loss of generality or overlooked auxiliary mechanisms.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Reward Score Matching unifies reward-based fine-tuning for flow and diffusion models by recasting alignment as score matching to a value-guided target.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Many reward-based fine-tuning methods for diffusion and flow models reduce to a single score-matching objective against a value-guided target.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"4a6e37e38fb316c83f3835af855adce2bbccf2923c582855d1481dae5a07005b"},"source":{"id":"2604.17415","kind":"arxiv","version":3},"verdict":{"id":"be8edd1a-49b9-46d3-ba94-a222182b8113","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T05:30:29.711746Z","strongest_claim":"Although existing methods are derived from different perspectives, we show that many can be written under a common framework, which we call reward score matching (RSM).","one_line_summary":"Reward Score Matching unifies reward-based fine-tuning for flow and diffusion models by recasting alignment as score matching to a value-guided target.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the primary distinctions among existing methods reduce to the construction of the value-guidance estimator and the effective optimization strength across timesteps, without material loss of generality or overlooked auxiliary mechanisms.","pith_extraction_headline":"Many reward-based fine-tuning methods for diffusion and flow models reduce to a single score-matching objective against a value-guided target."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.17415/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":"2604.17415","created_at":"2026-06-02T01:03:47.239651+00:00"},{"alias_kind":"arxiv_version","alias_value":"2604.17415v3","created_at":"2026-06-02T01:03:47.239651+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.17415","created_at":"2026-06-02T01:03:47.239651+00:00"},{"alias_kind":"pith_short_12","alias_value":"IOVTADSAIQNR","created_at":"2026-06-02T01:03:47.239651+00:00"},{"alias_kind":"pith_short_16","alias_value":"IOVTADSAIQNRQIZO","created_at":"2026-06-02T01:03:47.239651+00:00"},{"alias_kind":"pith_short_8","alias_value":"IOVTADSA","created_at":"2026-06-02T01:03:47.239651+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2605.06583","citing_title":"Improved techniques for fine-tuning flow models via adjoint matching: a deterministic control pipeline","ref_index":19,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IOVTADSAIQNRQIZOM5ICML5RTE","json":"https://pith.science/pith/IOVTADSAIQNRQIZOM5ICML5RTE.json","graph_json":"https://pith.science/api/pith-number/IOVTADSAIQNRQIZOM5ICML5RTE/graph.json","events_json":"https://pith.science/api/pith-number/IOVTADSAIQNRQIZOM5ICML5RTE/events.json","paper":"https://pith.science/paper/IOVTADSA"},"agent_actions":{"view_html":"https://pith.science/pith/IOVTADSAIQNRQIZOM5ICML5RTE","download_json":"https://pith.science/pith/IOVTADSAIQNRQIZOM5ICML5RTE.json","view_paper":"https://pith.science/paper/IOVTADSA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2604.17415&json=true","fetch_graph":"https://pith.science/api/pith-number/IOVTADSAIQNRQIZOM5ICML5RTE/graph.json","fetch_events":"https://pith.science/api/pith-number/IOVTADSAIQNRQIZOM5ICML5RTE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IOVTADSAIQNRQIZOM5ICML5RTE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IOVTADSAIQNRQIZOM5ICML5RTE/action/storage_attestation","attest_author":"https://pith.science/pith/IOVTADSAIQNRQIZOM5ICML5RTE/action/author_attestation","sign_citation":"https://pith.science/pith/IOVTADSAIQNRQIZOM5ICML5RTE/action/citation_signature","submit_replication":"https://pith.science/pith/IOVTADSAIQNRQIZOM5ICML5RTE/action/replication_record"}},"created_at":"2026-06-02T01:03:47.239651+00:00","updated_at":"2026-06-02T01:03:47.239651+00:00"}