{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:JEZOOPVV3ZDLTORKVIPSSWD3KB","short_pith_number":"pith:JEZOOPVV","schema_version":"1.0","canonical_sha256":"4932e73eb5de46b9ba2aaa1f29587b50764976fa8b621f644d369952245a61e4","source":{"kind":"arxiv","id":"2302.10688","version":3},"attestation_state":"computed","paper":{"title":"On Calibrating Diffusion Probabilistic Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Chao Du, Cheng Lu, Min Lin, Shuicheng Yan, Tianyu Pang, Zhijie Deng","submitted_at":"2023-02-21T14:14:40Z","abstract_excerpt":"Recently, diffusion probabilistic models (DPMs) have achieved promising results in diverse generative tasks. A typical DPM framework includes a forward process that gradually diffuses the data distribution and a reverse process that recovers the data distribution from time-dependent data scores. In this work, we observe that the stochastic reverse process of data scores is a martingale, from which concentration bounds and the optional stopping theorem for data scores can be derived. Then, we discover a simple way for calibrating an arbitrary pretrained DPM, with which the score matching loss c"},"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":"2302.10688","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-02-21T14:14:40Z","cross_cats_sorted":["cs.CV","stat.ML"],"title_canon_sha256":"41e839eb4aa0c3b8c1b81b85741de24b33faf4021eb962233984a89689bfa485","abstract_canon_sha256":"f0487506a27811fd93f448f28be1ed0a9ae71741fd1607392c9117e6bcd59b3c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:06:23.564568Z","signature_b64":"KrcqzHT/M8VKExFP8nSnhmTH8kfsom5xw+RB22MfwTjREaJOw9M10rj62H+yJoaoKD/0bYwOx26IyvGhAJYBAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4932e73eb5de46b9ba2aaa1f29587b50764976fa8b621f644d369952245a61e4","last_reissued_at":"2026-07-05T07:06:23.564054Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:06:23.564054Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"On Calibrating Diffusion Probabilistic Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Chao Du, Cheng Lu, Min Lin, Shuicheng Yan, Tianyu Pang, Zhijie Deng","submitted_at":"2023-02-21T14:14:40Z","abstract_excerpt":"Recently, diffusion probabilistic models (DPMs) have achieved promising results in diverse generative tasks. A typical DPM framework includes a forward process that gradually diffuses the data distribution and a reverse process that recovers the data distribution from time-dependent data scores. In this work, we observe that the stochastic reverse process of data scores is a martingale, from which concentration bounds and the optional stopping theorem for data scores can be derived. Then, we discover a simple way for calibrating an arbitrary pretrained DPM, with which the score matching loss c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2302.10688","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/2302.10688/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":"2302.10688","created_at":"2026-07-05T07:06:23.564111+00:00"},{"alias_kind":"arxiv_version","alias_value":"2302.10688v3","created_at":"2026-07-05T07:06:23.564111+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2302.10688","created_at":"2026-07-05T07:06:23.564111+00:00"},{"alias_kind":"pith_short_12","alias_value":"JEZOOPVV3ZDL","created_at":"2026-07-05T07:06:23.564111+00:00"},{"alias_kind":"pith_short_16","alias_value":"JEZOOPVV3ZDLTORK","created_at":"2026-07-05T07:06:23.564111+00:00"},{"alias_kind":"pith_short_8","alias_value":"JEZOOPVV","created_at":"2026-07-05T07:06:23.564111+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2509.18611","citing_title":"Flow marching for a generative PDE foundation model","ref_index":49,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JEZOOPVV3ZDLTORKVIPSSWD3KB","json":"https://pith.science/pith/JEZOOPVV3ZDLTORKVIPSSWD3KB.json","graph_json":"https://pith.science/api/pith-number/JEZOOPVV3ZDLTORKVIPSSWD3KB/graph.json","events_json":"https://pith.science/api/pith-number/JEZOOPVV3ZDLTORKVIPSSWD3KB/events.json","paper":"https://pith.science/paper/JEZOOPVV"},"agent_actions":{"view_html":"https://pith.science/pith/JEZOOPVV3ZDLTORKVIPSSWD3KB","download_json":"https://pith.science/pith/JEZOOPVV3ZDLTORKVIPSSWD3KB.json","view_paper":"https://pith.science/paper/JEZOOPVV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2302.10688&json=true","fetch_graph":"https://pith.science/api/pith-number/JEZOOPVV3ZDLTORKVIPSSWD3KB/graph.json","fetch_events":"https://pith.science/api/pith-number/JEZOOPVV3ZDLTORKVIPSSWD3KB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JEZOOPVV3ZDLTORKVIPSSWD3KB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JEZOOPVV3ZDLTORKVIPSSWD3KB/action/storage_attestation","attest_author":"https://pith.science/pith/JEZOOPVV3ZDLTORKVIPSSWD3KB/action/author_attestation","sign_citation":"https://pith.science/pith/JEZOOPVV3ZDLTORKVIPSSWD3KB/action/citation_signature","submit_replication":"https://pith.science/pith/JEZOOPVV3ZDLTORKVIPSSWD3KB/action/replication_record"}},"created_at":"2026-07-05T07:06:23.564111+00:00","updated_at":"2026-07-05T07:06:23.564111+00:00"}