{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:V7V2G2JRPWUDDH6NKUITWJ3QWD","short_pith_number":"pith:V7V2G2JR","schema_version":"1.0","canonical_sha256":"afeba369317da8319fcd55113b2770b0d000122891d145ad102b17b552370702","source":{"kind":"arxiv","id":"2101.09258","version":4},"attestation_state":"computed","paper":{"title":"Maximum Likelihood Training of Score-Based Diffusion Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Conor Durkan, Iain Murray, Stefano Ermon, Yang Song","submitted_at":"2021-01-22T18:22:29Z","abstract_excerpt":"Score-based diffusion models synthesize samples by reversing a stochastic process that diffuses data to noise, and are trained by minimizing a weighted combination of score matching losses. The log-likelihood of score-based diffusion models can be tractably computed through a connection to continuous normalizing flows, but log-likelihood is not directly optimized by the weighted combination of score matching losses. We show that for a specific weighting scheme, the objective upper bounds the negative log-likelihood, thus enabling approximate maximum likelihood training of score-based diffusion"},"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":"2101.09258","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2021-01-22T18:22:29Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"9a283eaf948783e34c9f5c0cc8333d054190a0f8f745156e8afd17ad87fb2407","abstract_canon_sha256":"dd3893fc014837797c4c63ccd08731844cf0bc3cd57fe1bf1a2fbb0ad16923d7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:24:28.106273Z","signature_b64":"Rt25JR+AeovcvKWMtGy3DO0hEthxe5LjpXuRnzY1EVXKFAQif1a/5LuMPT3f2V9dc41R2XNK6q2TzA5GzusBDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"afeba369317da8319fcd55113b2770b0d000122891d145ad102b17b552370702","last_reissued_at":"2026-07-05T03:24:28.105792Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:24:28.105792Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Maximum Likelihood Training of Score-Based Diffusion Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Conor Durkan, Iain Murray, Stefano Ermon, Yang Song","submitted_at":"2021-01-22T18:22:29Z","abstract_excerpt":"Score-based diffusion models synthesize samples by reversing a stochastic process that diffuses data to noise, and are trained by minimizing a weighted combination of score matching losses. The log-likelihood of score-based diffusion models can be tractably computed through a connection to continuous normalizing flows, but log-likelihood is not directly optimized by the weighted combination of score matching losses. We show that for a specific weighting scheme, the objective upper bounds the negative log-likelihood, thus enabling approximate maximum likelihood training of score-based diffusion"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2101.09258","kind":"arxiv","version":4},"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/2101.09258/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":"2101.09258","created_at":"2026-07-05T03:24:28.105851+00:00"},{"alias_kind":"arxiv_version","alias_value":"2101.09258v4","created_at":"2026-07-05T03:24:28.105851+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2101.09258","created_at":"2026-07-05T03:24:28.105851+00:00"},{"alias_kind":"pith_short_12","alias_value":"V7V2G2JRPWUD","created_at":"2026-07-05T03:24:28.105851+00:00"},{"alias_kind":"pith_short_16","alias_value":"V7V2G2JRPWUDDH6N","created_at":"2026-07-05T03:24:28.105851+00:00"},{"alias_kind":"pith_short_8","alias_value":"V7V2G2JR","created_at":"2026-07-05T03:24:28.105851+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2602.11229","citing_title":"Latent Generative Solvers for Generalizable Long-Term Physics Simulation","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07907","citing_title":"Consistency Regularised Gradient Flows for Inverse Problems","ref_index":60,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/V7V2G2JRPWUDDH6NKUITWJ3QWD","json":"https://pith.science/pith/V7V2G2JRPWUDDH6NKUITWJ3QWD.json","graph_json":"https://pith.science/api/pith-number/V7V2G2JRPWUDDH6NKUITWJ3QWD/graph.json","events_json":"https://pith.science/api/pith-number/V7V2G2JRPWUDDH6NKUITWJ3QWD/events.json","paper":"https://pith.science/paper/V7V2G2JR"},"agent_actions":{"view_html":"https://pith.science/pith/V7V2G2JRPWUDDH6NKUITWJ3QWD","download_json":"https://pith.science/pith/V7V2G2JRPWUDDH6NKUITWJ3QWD.json","view_paper":"https://pith.science/paper/V7V2G2JR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2101.09258&json=true","fetch_graph":"https://pith.science/api/pith-number/V7V2G2JRPWUDDH6NKUITWJ3QWD/graph.json","fetch_events":"https://pith.science/api/pith-number/V7V2G2JRPWUDDH6NKUITWJ3QWD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/V7V2G2JRPWUDDH6NKUITWJ3QWD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/V7V2G2JRPWUDDH6NKUITWJ3QWD/action/storage_attestation","attest_author":"https://pith.science/pith/V7V2G2JRPWUDDH6NKUITWJ3QWD/action/author_attestation","sign_citation":"https://pith.science/pith/V7V2G2JRPWUDDH6NKUITWJ3QWD/action/citation_signature","submit_replication":"https://pith.science/pith/V7V2G2JRPWUDDH6NKUITWJ3QWD/action/replication_record"}},"created_at":"2026-07-05T03:24:28.105851+00:00","updated_at":"2026-07-05T03:24:28.105851+00:00"}