{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:JZBKGJFH5IEGBFWM2NI6HWGPBD","short_pith_number":"pith:JZBKGJFH","schema_version":"1.0","canonical_sha256":"4e42a324a7ea086096ccd351e3d8cf08f95dda2bdfe203a5729316100792200f","source":{"kind":"arxiv","id":"2602.05533","version":3},"attestation_state":"computed","paper":{"title":"Conditional Diffusion Guidance under Hard Constraint: A Stochastic Analysis Approach","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Renyuan Xu, Wenpin Tang, Zhengyi Guo","submitted_at":"2026-02-05T10:46:20Z","abstract_excerpt":"We study conditional generation in diffusion models under hard constraints, where generated samples must satisfy prescribed events with probability one. Such constraints arise naturally in safety-critical applications and in rare-event simulation, where soft or reward-based guidance methods offer no guarantee of constraint satisfaction. Building on a probabilistic interpretation of diffusion models, we develop a principled conditional diffusion guidance framework based on Doob's h-transform, martingale representation and quadratic variation process. Specifically, the resulting guided dynamics "},"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":"2602.05533","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-02-05T10:46:20Z","cross_cats_sorted":[],"title_canon_sha256":"0e140f189e4947de775845c89c377ae1355da47dd63190429cda386e3313a218","abstract_canon_sha256":"c14cf1d64cd5bb26c2cb774960f4684fb51f2bb49fc84f17801d58a1c2ed698f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-19T16:12:18.482098Z","signature_b64":"7Sm81oXso66R30/1sPgtu6RbI9LGybUMMemb5x32JdxNjTTo6RjFJ+ocPMKvoJXnzBbrelgokJ07Vs1MMkiXBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4e42a324a7ea086096ccd351e3d8cf08f95dda2bdfe203a5729316100792200f","last_reissued_at":"2026-06-19T16:12:18.481728Z","signature_status":"signed_v1","first_computed_at":"2026-06-19T16:12:18.481728Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Conditional Diffusion Guidance under Hard Constraint: A Stochastic Analysis Approach","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Renyuan Xu, Wenpin Tang, Zhengyi Guo","submitted_at":"2026-02-05T10:46:20Z","abstract_excerpt":"We study conditional generation in diffusion models under hard constraints, where generated samples must satisfy prescribed events with probability one. Such constraints arise naturally in safety-critical applications and in rare-event simulation, where soft or reward-based guidance methods offer no guarantee of constraint satisfaction. Building on a probabilistic interpretation of diffusion models, we develop a principled conditional diffusion guidance framework based on Doob's h-transform, martingale representation and quadratic variation process. Specifically, the resulting guided dynamics "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.05533","kind":"arxiv","version":3},"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/2602.05533/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":"2602.05533","created_at":"2026-06-19T16:12:18.481775+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.05533v3","created_at":"2026-06-19T16:12:18.481775+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.05533","created_at":"2026-06-19T16:12:18.481775+00:00"},{"alias_kind":"pith_short_12","alias_value":"JZBKGJFH5IEG","created_at":"2026-06-19T16:12:18.481775+00:00"},{"alias_kind":"pith_short_16","alias_value":"JZBKGJFH5IEGBFWM","created_at":"2026-06-19T16:12:18.481775+00:00"},{"alias_kind":"pith_short_8","alias_value":"JZBKGJFH","created_at":"2026-06-19T16:12:18.481775+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":6,"sample":[{"citing_arxiv_id":"2605.07220","citing_title":"On the Robustness of Distribution Support under Diffusion Guidance","ref_index":57,"is_internal_anchor":true},{"citing_arxiv_id":"2605.18719","citing_title":"SafeDiffusion-R1: Online Reward Steering for Safe Diffusion Post-Training","ref_index":123,"is_internal_anchor":true},{"citing_arxiv_id":"2605.19317","citing_title":"Inference-Time Scaling in Diffusion Models through Iterative Partial Refinement","ref_index":7,"is_internal_anchor":true},{"citing_arxiv_id":"2605.19391","citing_title":"Tweedie's Formulae and Diffusion Generative Models Beyond Gaussian","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2605.06538","citing_title":"Diffusion-Based Posterior Sampling: A Feynman-Kac Analysis of Bias and Stability","ref_index":9,"is_internal_anchor":true},{"citing_arxiv_id":"2605.07220","citing_title":"On the Robustness of Distribution Support under Diffusion Guidance","ref_index":57,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JZBKGJFH5IEGBFWM2NI6HWGPBD","json":"https://pith.science/pith/JZBKGJFH5IEGBFWM2NI6HWGPBD.json","graph_json":"https://pith.science/api/pith-number/JZBKGJFH5IEGBFWM2NI6HWGPBD/graph.json","events_json":"https://pith.science/api/pith-number/JZBKGJFH5IEGBFWM2NI6HWGPBD/events.json","paper":"https://pith.science/paper/JZBKGJFH"},"agent_actions":{"view_html":"https://pith.science/pith/JZBKGJFH5IEGBFWM2NI6HWGPBD","download_json":"https://pith.science/pith/JZBKGJFH5IEGBFWM2NI6HWGPBD.json","view_paper":"https://pith.science/paper/JZBKGJFH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.05533&json=true","fetch_graph":"https://pith.science/api/pith-number/JZBKGJFH5IEGBFWM2NI6HWGPBD/graph.json","fetch_events":"https://pith.science/api/pith-number/JZBKGJFH5IEGBFWM2NI6HWGPBD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JZBKGJFH5IEGBFWM2NI6HWGPBD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JZBKGJFH5IEGBFWM2NI6HWGPBD/action/storage_attestation","attest_author":"https://pith.science/pith/JZBKGJFH5IEGBFWM2NI6HWGPBD/action/author_attestation","sign_citation":"https://pith.science/pith/JZBKGJFH5IEGBFWM2NI6HWGPBD/action/citation_signature","submit_replication":"https://pith.science/pith/JZBKGJFH5IEGBFWM2NI6HWGPBD/action/replication_record"}},"created_at":"2026-06-19T16:12:18.481775+00:00","updated_at":"2026-06-19T16:12:18.481775+00:00"}