{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:4DNWNNUOUCIBV7Q33NXOYRNZIA","short_pith_number":"pith:4DNWNNUO","schema_version":"1.0","canonical_sha256":"e0db66b68ea0901afe1bdb6eec45b94005bfbf35048d2c1604061cd4edd41aa1","source":{"kind":"arxiv","id":"2605.11246","version":2},"attestation_state":"computed","paper":{"title":"Support-Proximity Augmented Diffusion Estimation for Offline Black-Box Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"SPADE augments diffusion models with support-proximity regularization to solve offline black-box optimization more effectively.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Bowei He, Can Chen, Haolun Wu, Linfeng Du, Xue Liu, Ye Yuan, Yonghan Yang, Zipeng Sun","submitted_at":"2026-05-11T21:09:28Z","abstract_excerpt":"Offline black-box optimization aims to discover novel designs with high property scores using only a static dataset, a task fundamentally challenged by the out-of-distribution (OOD) extrapolation problem. Existing approaches typically bifurcate into inverse methods, which struggle with the ill-posed nature of mapping scores to designs, and forward methods, which often lack the distributional expressivity to quantify uncertainty effectively. In this work, we propose SPADE (Support-Proximity Augmented Diffusion Estimation), a novel framework that reimagines forward surrogate modeling through 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":true},"canonical_record":{"source":{"id":"2605.11246","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-11T21:09:28Z","cross_cats_sorted":[],"title_canon_sha256":"3543e9f49a00ac3d720c581bd200e75a00137da6284da5d14904ec3f6b24a067","abstract_canon_sha256":"9cb2af30a39151a7801f3a03c58f4042e468bded850a9daa4efdc575143cdb7c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-22T01:04:55.439306Z","signature_b64":"ZllZnLvZn/FU7cADMfKEI2e6vyT3UjECi6oyQ2RFVW3mpUkpc5DqhD/cxqX91vXCjo+orgtoyOCtR/lo+gN0Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e0db66b68ea0901afe1bdb6eec45b94005bfbf35048d2c1604061cd4edd41aa1","last_reissued_at":"2026-05-22T01:04:55.438509Z","signature_status":"signed_v1","first_computed_at":"2026-05-22T01:04:55.438509Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Support-Proximity Augmented Diffusion Estimation for Offline Black-Box Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"SPADE augments diffusion models with support-proximity regularization to solve offline black-box optimization more effectively.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Bowei He, Can Chen, Haolun Wu, Linfeng Du, Xue Liu, Ye Yuan, Yonghan Yang, Zipeng Sun","submitted_at":"2026-05-11T21:09:28Z","abstract_excerpt":"Offline black-box optimization aims to discover novel designs with high property scores using only a static dataset, a task fundamentally challenged by the out-of-distribution (OOD) extrapolation problem. Existing approaches typically bifurcate into inverse methods, which struggle with the ill-posed nature of mapping scores to designs, and forward methods, which often lack the distributional expressivity to quantify uncertainty effectively. In this work, we propose SPADE (Support-Proximity Augmented Diffusion Estimation), a novel framework that reimagines forward surrogate modeling through the"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"SPADE achieves state-of-the-art performance across Design-Bench tasks and an LLM data mixture optimization benchmark. Theoretically, we prove that our regularization is first-order equivalent to maximizing a Bayesian posterior with a valid design prior.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That kNN-based density estimation on the input space sufficiently captures the data manifold constraint p(x) and that the added calibration module can enforce global moment and ranking consistency without distorting the learned conditional distribution p(y|x).","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"SPADE augments conditional diffusion models for forward surrogate modeling in offline BBO with calibrated moment/ranking consistency and support-proximity regularization, achieving SOTA on Design-Bench and LLM optimization benchmarks.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"SPADE augments diffusion models with support-proximity regularization to solve offline black-box optimization more effectively.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"cbbaeedbd912b657ef7e917d215d4329256fe5248d9f41574c866616a8efd951"},"source":{"id":"2605.11246","kind":"arxiv","version":2},"verdict":{"id":"a59c22b6-7d37-4023-ba24-0184b2b0d498","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-13T01:58:31.179220Z","strongest_claim":"SPADE achieves state-of-the-art performance across Design-Bench tasks and an LLM data mixture optimization benchmark. Theoretically, we prove that our regularization is first-order equivalent to maximizing a Bayesian posterior with a valid design prior.","one_line_summary":"SPADE augments conditional diffusion models for forward surrogate modeling in offline BBO with calibrated moment/ranking consistency and support-proximity regularization, achieving SOTA on Design-Bench and LLM optimization benchmarks.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That kNN-based density estimation on the input space sufficiently captures the data manifold constraint p(x) and that the added calibration module can enforce global moment and ranking consistency without distorting the learned conditional distribution p(y|x).","pith_extraction_headline":"SPADE augments diffusion models with support-proximity regularization to solve offline black-box optimization more effectively."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.11246/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"claim_evidence","ran_at":"2026-05-20T04:42:00.835326Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T12:39:46.424720Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T10:01:17.363565Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T08:38:05.572418Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"c20c3de36045d14423d1c0af4f746cb501c4565f62eb7837be734651c99a3698"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":1,"snapshot_sha256":"e36ac71ca014f2d6114010a838b56efc70d2548977972550cb7424e10a1b7b46"},"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":"2605.11246","created_at":"2026-05-22T01:04:55.438631+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.11246v2","created_at":"2026-05-22T01:04:55.438631+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.11246","created_at":"2026-05-22T01:04:55.438631+00:00"},{"alias_kind":"pith_short_12","alias_value":"4DNWNNUOUCIB","created_at":"2026-05-22T01:04:55.438631+00:00"},{"alias_kind":"pith_short_16","alias_value":"4DNWNNUOUCIBV7Q3","created_at":"2026-05-22T01:04:55.438631+00:00"},{"alias_kind":"pith_short_8","alias_value":"4DNWNNUO","created_at":"2026-05-22T01:04:55.438631+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2605.22144","citing_title":"One Sentence, One Drama: Personalized Short-Form Drama Generation via Multi-Agent Systems","ref_index":50,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":1,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4DNWNNUOUCIBV7Q33NXOYRNZIA","json":"https://pith.science/pith/4DNWNNUOUCIBV7Q33NXOYRNZIA.json","graph_json":"https://pith.science/api/pith-number/4DNWNNUOUCIBV7Q33NXOYRNZIA/graph.json","events_json":"https://pith.science/api/pith-number/4DNWNNUOUCIBV7Q33NXOYRNZIA/events.json","paper":"https://pith.science/paper/4DNWNNUO"},"agent_actions":{"view_html":"https://pith.science/pith/4DNWNNUOUCIBV7Q33NXOYRNZIA","download_json":"https://pith.science/pith/4DNWNNUOUCIBV7Q33NXOYRNZIA.json","view_paper":"https://pith.science/paper/4DNWNNUO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.11246&json=true","fetch_graph":"https://pith.science/api/pith-number/4DNWNNUOUCIBV7Q33NXOYRNZIA/graph.json","fetch_events":"https://pith.science/api/pith-number/4DNWNNUOUCIBV7Q33NXOYRNZIA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4DNWNNUOUCIBV7Q33NXOYRNZIA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4DNWNNUOUCIBV7Q33NXOYRNZIA/action/storage_attestation","attest_author":"https://pith.science/pith/4DNWNNUOUCIBV7Q33NXOYRNZIA/action/author_attestation","sign_citation":"https://pith.science/pith/4DNWNNUOUCIBV7Q33NXOYRNZIA/action/citation_signature","submit_replication":"https://pith.science/pith/4DNWNNUOUCIBV7Q33NXOYRNZIA/action/replication_record"}},"created_at":"2026-05-22T01:04:55.438631+00:00","updated_at":"2026-05-22T01:04:55.438631+00:00"}