{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:425FAKATCEINVWPTAU2YBOM7E4","merge_version":"pith-open-graph-merge-v1","event_count":4,"valid_event_count":4,"invalid_event_count":0,"equivocation_count":1,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"db317580ce80f391c0eacf843fd6d872934d20ad2b1df51c1c7cefa88b3d1628","cross_cats_sorted":["cs.CV","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2026-05-13T07:46:06Z","title_canon_sha256":"cf0392d79de7e1d5f4e6f7649d50b3c88c6ead065fc6365eb9586b3a19714a62"},"schema_version":"1.0","source":{"id":"2605.13910","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13910","created_at":"2026-05-17T23:39:18Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13910v1","created_at":"2026-05-17T23:39:18Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13910","created_at":"2026-05-17T23:39:18Z"},{"alias_kind":"pith_short_12","alias_value":"425FAKATCEIN","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"425FAKATCEINVWPT","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"425FAKAT","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:078ee174dd087e0caf26892e89fc8b3825fb3b8c63deed0cd3c32e80d3624270","target":"graph","created_at":"2026-05-17T23:39:18Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"For pixel-based DMs, our method consistently produces superior samples compared to state-of-the-art second order samplers (Heun, DPM-Solver++) and the recent aDDIM sampler, at an identical number of function evaluations (NFE)."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The hypothesis that samplers fail in the few-step regime solely because they rely only on the predicted mean of the reverse distribution, and that explicitly modeling the covariance via Tweedie's formula plus Fourier decomposition will reliably fix it without introducing new instabilities."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"A covariance-aware extension of DDIM sampling for pixel-space diffusion models that uses Tweedie's formula and Fourier decomposition to model reverse-process covariance and improves sample quality at low NFE."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Modeling the full reverse-process covariance improves few-step sampling in pixel-space diffusion models."}],"snapshot_sha256":"65754af440d9b5c400b66ac1b33577fe4ded8f761685747b789789f010931a82"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"24abca46e0e175e1d5e41f02d87c93666a887eef3fad6c2d40c62835bde1cb77"},"paper":{"abstract_excerpt":"We present a covariance-aware sampler that improves the quality of pixel-space Diffusion Model (DM) sampling in the few-step regime. We hypothesize that in the few-step regime samplers fail because they rely solely on the predicted mean of the reverse distribution, while our solution explicitly models the reverse-process covariance. Our method combines Tweedie's formula to estimate the covariance with an efficient, structured Fourier-space decomposition of the covariance matrix. Implemented as an extension of DDIM, our method requires only a minimal overhead: one extra Jacobian-Vector Product ","authors_text":"Andrea Schioppa, Tim Salimans","cross_cats":["cs.CV","cs.LG"],"headline":"Modeling the full reverse-process covariance improves few-step sampling in pixel-space diffusion models.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2026-05-13T07:46:06Z","title":"Covariance-aware sampling for Diffusion Models"},"references":{"count":22,"internal_anchors":2,"resolved_work":22,"sample":[{"cited_arxiv_id":"","doi":"10.1109/t-c.1974.223784","is_internal_anchor":false,"ref_index":1,"title":"doi: 10.1109/T-C.1974.223784. A. Blattmann, T. Dockhorn, S. Kulal, D. Mendelevitch, M. Kilian, D. Lorenz, Y. Levi, Z. English, V. Voleti, A. Letts, et al. Stable video diffusion: Scaling latent video ","work_id":"37396299-0db1-4fcd-9202-a7e69695fb3a","year":1974},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Tweedie moment projected diffusions for inverse problems","work_id":"dde56871-3be6-486f-9070-873b6114e4d9","year":2009},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"URLhttps://sander.ai/2024/09/02/ spectral-autoregression.html. T. Dockhorn, A. Vahdat, and K. Kreis. Genie: Higher-order denoising diffusion solvers. In S. Koyejo, S. Mohamed, A. Agarwal, D. Belgrave,","work_id":"917c905f-c803-48d5-a4c0-ffdd2d1ef884","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"URL https://proceedings.neurips.cc/paper_files/paper/2022/file/ c281c5a17ad2e55e1ac1ca825071f991-Paper-Conference.pdf. B. Efron. Tweedie’s formula and selection bias.Journal of the American Statistica","work_id":"4828cccc-a79b-4f37-be81-06e04f16cafb","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Tweedie’s Formula and Selection Bias.Journal of the American Statistical Association, 106(496):1602–1614","work_id":"025e78f0-da85-48fd-891d-f69ee62f2017","year":null}],"snapshot_sha256":"3e53692cb75d27d6ae80905e75e587fe0346a477d6d41dffc845ab865564d39b"},"source":{"id":"2605.13910","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-15T03:02:49.803340Z","id":"7c84ebad-b29b-418e-8ed9-87a18006d827","model_set":{"reader":"grok-4.3"},"one_line_summary":"A covariance-aware extension of DDIM sampling for pixel-space diffusion models that uses Tweedie's formula and Fourier decomposition to model reverse-process covariance and improves sample quality at low NFE.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Modeling the full reverse-process covariance improves few-step sampling in pixel-space diffusion models.","strongest_claim":"For pixel-based DMs, our method consistently produces superior samples compared to state-of-the-art second order samplers (Heun, DPM-Solver++) and the recent aDDIM sampler, at an identical number of function evaluations (NFE).","weakest_assumption":"The hypothesis that samplers fail in the few-step regime solely because they rely only on the predicted mean of the reverse distribution, and that explicitly modeling the covariance via Tweedie's formula plus Fourier decomposition will reliably fix it without introducing new instabilities."}},"verdict_id":"7c84ebad-b29b-418e-8ed9-87a18006d827"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:81f376a9cb076e11b8f463ae9e5de916de91743fd57636dc7eb7911309ed9099","target":"record","created_at":"2026-05-17T23:39:18Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"db317580ce80f391c0eacf843fd6d872934d20ad2b1df51c1c7cefa88b3d1628","cross_cats_sorted":["cs.CV","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2026-05-13T07:46:06Z","title_canon_sha256":"cf0392d79de7e1d5f4e6f7649d50b3c88c6ead065fc6365eb9586b3a19714a62"},"schema_version":"1.0","source":{"id":"2605.13910","kind":"arxiv","version":1}},"canonical_sha256":"e6ba5028131110dad9f3053580b99f272ca257b9452f50fd6c0d38b931e65e75","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e6ba5028131110dad9f3053580b99f272ca257b9452f50fd6c0d38b931e65e75","first_computed_at":"2026-05-17T23:39:18.824218Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:39:18.824218Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"UhE+rJ8O/nI3+DoETRp9uiF3mp2/lripBmoLIRvmVhj2c6Dxlx9pJ1GoFLEqXocx3Z60ZJAJWwPsCA7Du4qGBA==","signature_status":"signed_v1","signed_at":"2026-05-17T23:39:18.824941Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.13910","source_kind":"arxiv","source_version":1}}},"equivocations":[{"signer_id":"pith.science","event_type":"integrity_finding","target":"integrity","event_ids":["sha256:4ddcd9be324feb965299525e633be5e6e82a89d67f1a9be3ef283b95d5f682cf","sha256:fb1444dcf73894c2f9131b2b98231afa8cff58bf9ee7af735cf3ac968ec5ed69"]}],"invalid_events":[],"applied_event_ids":["sha256:81f376a9cb076e11b8f463ae9e5de916de91743fd57636dc7eb7911309ed9099","sha256:078ee174dd087e0caf26892e89fc8b3825fb3b8c63deed0cd3c32e80d3624270"],"state_sha256":"7ec0db879e391ad48fcea89de82bb21d76dd5bed900fc9911974872f467c06ec"}