{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:LJQKASFUH4NGJ46OXCPU3MFNRN","short_pith_number":"pith:LJQKASFU","schema_version":"1.0","canonical_sha256":"5a60a048b43f1a64f3ceb89f4db0ad8b58a268d5d477b4c03e3e1c65d587da7a","source":{"kind":"arxiv","id":"2406.04103","version":1},"attestation_state":"computed","paper":{"title":"Multistep Distillation of Diffusion Models via Moment Matching","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.NE"],"primary_cat":"cs.LG","authors_text":"Emiel Hoogeboom, Jonathan Heek, Thomas Mensink, Tim Salimans","submitted_at":"2024-06-06T14:20:21Z","abstract_excerpt":"We present a new method for making diffusion models faster to sample. The method distills many-step diffusion models into few-step models by matching conditional expectations of the clean data given noisy data along the sampling trajectory. Our approach extends recently proposed one-step methods to the multi-step case, and provides a new perspective by interpreting these approaches in terms of moment matching. By using up to 8 sampling steps, we obtain distilled models that outperform not only their one-step versions but also their original many-step teacher models, obtaining new state-of-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":"2406.04103","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2024-06-06T14:20:21Z","cross_cats_sorted":["cs.AI","cs.CV","cs.NE"],"title_canon_sha256":"950f878957bdecf249c4f81b2b11a99701fdccf7eeecf57d9b5cbc299a63040e","abstract_canon_sha256":"20a1b164272567a53c0cce9a9d1adb788b2798a2f6569a41fc70ff7007433942"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:28:20.784463Z","signature_b64":"NOHdSYMeajFKfxyGh8yDWhmiosEUaSXE8WngijX67LukTLaIr1b4yWXpTpehHKHeGwRpb2hPwdpqOGBwzf7ZDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5a60a048b43f1a64f3ceb89f4db0ad8b58a268d5d477b4c03e3e1c65d587da7a","last_reissued_at":"2026-07-05T08:28:20.783846Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:28:20.783846Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Multistep Distillation of Diffusion Models via Moment Matching","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.NE"],"primary_cat":"cs.LG","authors_text":"Emiel Hoogeboom, Jonathan Heek, Thomas Mensink, Tim Salimans","submitted_at":"2024-06-06T14:20:21Z","abstract_excerpt":"We present a new method for making diffusion models faster to sample. The method distills many-step diffusion models into few-step models by matching conditional expectations of the clean data given noisy data along the sampling trajectory. Our approach extends recently proposed one-step methods to the multi-step case, and provides a new perspective by interpreting these approaches in terms of moment matching. By using up to 8 sampling steps, we obtain distilled models that outperform not only their one-step versions but also their original many-step teacher models, obtaining new state-of-the-"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.04103","kind":"arxiv","version":1},"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/2406.04103/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":"2406.04103","created_at":"2026-07-05T08:28:20.783928+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.04103v1","created_at":"2026-07-05T08:28:20.783928+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.04103","created_at":"2026-07-05T08:28:20.783928+00:00"},{"alias_kind":"pith_short_12","alias_value":"LJQKASFUH4NG","created_at":"2026-07-05T08:28:20.783928+00:00"},{"alias_kind":"pith_short_16","alias_value":"LJQKASFUH4NGJ46O","created_at":"2026-07-05T08:28:20.783928+00:00"},{"alias_kind":"pith_short_8","alias_value":"LJQKASFU","created_at":"2026-07-05T08:28:20.783928+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.21489","citing_title":"Variance Reduction for Expectations with Diffusion Teachers","ref_index":68,"is_internal_anchor":false},{"citing_arxiv_id":"2412.15689","citing_title":"DOLLAR: Few-Step Video Generation via Distillation and Latent Reward Optimization","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21489","citing_title":"Variance Reduction for Expectations with Diffusion Teachers","ref_index":68,"is_internal_anchor":false},{"citing_arxiv_id":"2602.05449","citing_title":"DisCa: Accelerating Video Diffusion Transformers with Distillation-Compatible Learnable Feature Caching","ref_index":55,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LJQKASFUH4NGJ46OXCPU3MFNRN","json":"https://pith.science/pith/LJQKASFUH4NGJ46OXCPU3MFNRN.json","graph_json":"https://pith.science/api/pith-number/LJQKASFUH4NGJ46OXCPU3MFNRN/graph.json","events_json":"https://pith.science/api/pith-number/LJQKASFUH4NGJ46OXCPU3MFNRN/events.json","paper":"https://pith.science/paper/LJQKASFU"},"agent_actions":{"view_html":"https://pith.science/pith/LJQKASFUH4NGJ46OXCPU3MFNRN","download_json":"https://pith.science/pith/LJQKASFUH4NGJ46OXCPU3MFNRN.json","view_paper":"https://pith.science/paper/LJQKASFU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.04103&json=true","fetch_graph":"https://pith.science/api/pith-number/LJQKASFUH4NGJ46OXCPU3MFNRN/graph.json","fetch_events":"https://pith.science/api/pith-number/LJQKASFUH4NGJ46OXCPU3MFNRN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LJQKASFUH4NGJ46OXCPU3MFNRN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LJQKASFUH4NGJ46OXCPU3MFNRN/action/storage_attestation","attest_author":"https://pith.science/pith/LJQKASFUH4NGJ46OXCPU3MFNRN/action/author_attestation","sign_citation":"https://pith.science/pith/LJQKASFUH4NGJ46OXCPU3MFNRN/action/citation_signature","submit_replication":"https://pith.science/pith/LJQKASFUH4NGJ46OXCPU3MFNRN/action/replication_record"}},"created_at":"2026-07-05T08:28:20.783928+00:00","updated_at":"2026-07-05T08:28:20.783928+00:00"}