{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:WDFVIT2MXEEX7G2ZJ2MWIDGAQQ","short_pith_number":"pith:WDFVIT2M","schema_version":"1.0","canonical_sha256":"b0cb544f4cb9097f9b594e99640cc08406dff47b0a20d5db3e47fbbe9b24e05a","source":{"kind":"arxiv","id":"2605.15859","version":1},"attestation_state":"computed","paper":{"title":"Complexity of Non-Log-Concave Sampling in Fisher Information","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Leveraging log-concave sampling results gives non-log-concave sampling the same dimension dependence in relative Fisher information.","cross_cats":["cs.LG","math.ST","stat.ML","stat.TH"],"primary_cat":"cs.DS","authors_text":"Andre Wibisono, Sinho Chewi","submitted_at":"2026-05-15T11:20:26Z","abstract_excerpt":"We study the query complexity of obtaining a relative Fisher information guarantee for sampling from a log-smooth non-log-concave distribution; this is a sampling analog of finding an approximate stationary point in optimization. Our algorithm is based on the proximal sampler, which is an implicit discretization of the Langevin diffusion, and requires an implementation of the backward step known as the restricted Gaussian oracle (RGO). We show that by leveraging the recent results for log-concave sampling with high-accuracy guarantees in R\\'enyi divergence, we can obtain an approximate RGO imp"},"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":true,"formal_links_present":false},"canonical_record":{"source":{"id":"2605.15859","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.DS","submitted_at":"2026-05-15T11:20:26Z","cross_cats_sorted":["cs.LG","math.ST","stat.ML","stat.TH"],"title_canon_sha256":"12dec7eefd5adec411279f1dbb8370d041400fb5aa13f03e6c3c2de962692ed8","abstract_canon_sha256":"5324568a20a39bcf99cf560e6299ea2cbef29fe9c4bf5da405fb6d681f8deead"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:01:22.344311Z","signature_b64":"XfDQv+kuy/tPm7yswOJoRi+PA8OOlivIE2FGOyhTzAm9CYEveKD1MLe97NB3NtRE2RrQnwFsw9PxlzQihaMNCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b0cb544f4cb9097f9b594e99640cc08406dff47b0a20d5db3e47fbbe9b24e05a","last_reissued_at":"2026-05-20T00:01:22.343170Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:01:22.343170Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Complexity of Non-Log-Concave Sampling in Fisher Information","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Leveraging log-concave sampling results gives non-log-concave sampling the same dimension dependence in relative Fisher information.","cross_cats":["cs.LG","math.ST","stat.ML","stat.TH"],"primary_cat":"cs.DS","authors_text":"Andre Wibisono, Sinho Chewi","submitted_at":"2026-05-15T11:20:26Z","abstract_excerpt":"We study the query complexity of obtaining a relative Fisher information guarantee for sampling from a log-smooth non-log-concave distribution; this is a sampling analog of finding an approximate stationary point in optimization. Our algorithm is based on the proximal sampler, which is an implicit discretization of the Langevin diffusion, and requires an implementation of the backward step known as the restricted Gaussian oracle (RGO). We show that by leveraging the recent results for log-concave sampling with high-accuracy guarantees in R\\'enyi divergence, we can obtain an approximate RGO imp"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"by leveraging the recent results for log-concave sampling with high-accuracy guarantees in Rényi divergence, we can obtain an approximate RGO implementation that -- when used with the proximal sampler -- yields a complexity guarantee in relative Fisher information that inherits the same dimension dependence as log-concave sampling, and improves upon prior work for non-log-concave sampling.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The recent high-accuracy log-concave sampling results in Rényi divergence can be used to produce an approximate restricted Gaussian oracle whose error does not introduce worse dimension dependence when plugged into the proximal sampler for non-log-concave targets (abstract, paragraph describing the algorithm and RGO implementation).","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Proximal sampler with approximate RGO from log-concave Renyi results yields relative Fisher information complexity for non-log-concave sampling that matches log-concave dimension dependence, plus a converse reduction.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Leveraging log-concave sampling results gives non-log-concave sampling the same dimension dependence in relative Fisher information.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"1051f930a83682f40cb5927180543f5617424d0d5505c57e2755291bc6fa14e6"},"source":{"id":"2605.15859","kind":"arxiv","version":1},"verdict":{"id":"31522c2e-8327-438d-b95a-0f01b0674944","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T19:05:49.094374Z","strongest_claim":"by leveraging the recent results for log-concave sampling with high-accuracy guarantees in Rényi divergence, we can obtain an approximate RGO implementation that -- when used with the proximal sampler -- yields a complexity guarantee in relative Fisher information that inherits the same dimension dependence as log-concave sampling, and improves upon prior work for non-log-concave sampling.","one_line_summary":"Proximal sampler with approximate RGO from log-concave Renyi results yields relative Fisher information complexity for non-log-concave sampling that matches log-concave dimension dependence, plus a converse reduction.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The recent high-accuracy log-concave sampling results in Rényi divergence can be used to produce an approximate restricted Gaussian oracle whose error does not introduce worse dimension dependence when plugged into the proximal sampler for non-log-concave targets (abstract, paragraph describing the algorithm and RGO implementation).","pith_extraction_headline":"Leveraging log-concave sampling results gives non-log-concave sampling the same dimension dependence in relative Fisher information."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.15859/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-19T19:31:19.082948Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T19:12:36.049646Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T17:33:48.701982Z","status":"skipped","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T17:01:55.819740Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"d8d1aac0027c96405b5bfb3427dfeb9095f044d5ca4dfde08a5ca7b51c276bed"},"references":{"count":32,"sample":[{"doi":"","year":2022,"title":"and Salim, Adil and Zhang, Matthew S","work_id":"0961f6bd-aa96-464c-9787-d52220c57319","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Optimal score estimation via empirical","work_id":"b953e42c-0e74-40e7-a41c-ba3fd8b97316","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2026,"title":"Wibisono, Andre , journal=. 2026 , volume=","work_id":"924701c8-fbb9-4b27-ba2a-23c32d9ad2f2","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Sinho Chewi and Alkis Kalavasis and Anay Mehrotra and Omar Montasser , year=","work_id":"51182a8f-035f-4089-8e5a-e6d2f71e4316","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Proceedings of the 34th International Conference on Algorithmic Learning Theory , pages =","work_id":"4fbcc128-7679-4a5d-86a6-7c9a6dd24e93","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":32,"snapshot_sha256":"c3d5d4feb4e47cdbc6aa2e1ea01eb724cc54c36e015818da66d4d41d4a4260bb","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":"2605.15859","created_at":"2026-05-20T00:01:22.343309+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.15859v1","created_at":"2026-05-20T00:01:22.343309+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.15859","created_at":"2026-05-20T00:01:22.343309+00:00"},{"alias_kind":"pith_short_12","alias_value":"WDFVIT2MXEEX","created_at":"2026-05-20T00:01:22.343309+00:00"},{"alias_kind":"pith_short_16","alias_value":"WDFVIT2MXEEX7G2Z","created_at":"2026-05-20T00:01:22.343309+00:00"},{"alias_kind":"pith_short_8","alias_value":"WDFVIT2M","created_at":"2026-05-20T00:01:22.343309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WDFVIT2MXEEX7G2ZJ2MWIDGAQQ","json":"https://pith.science/pith/WDFVIT2MXEEX7G2ZJ2MWIDGAQQ.json","graph_json":"https://pith.science/api/pith-number/WDFVIT2MXEEX7G2ZJ2MWIDGAQQ/graph.json","events_json":"https://pith.science/api/pith-number/WDFVIT2MXEEX7G2ZJ2MWIDGAQQ/events.json","paper":"https://pith.science/paper/WDFVIT2M"},"agent_actions":{"view_html":"https://pith.science/pith/WDFVIT2MXEEX7G2ZJ2MWIDGAQQ","download_json":"https://pith.science/pith/WDFVIT2MXEEX7G2ZJ2MWIDGAQQ.json","view_paper":"https://pith.science/paper/WDFVIT2M","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.15859&json=true","fetch_graph":"https://pith.science/api/pith-number/WDFVIT2MXEEX7G2ZJ2MWIDGAQQ/graph.json","fetch_events":"https://pith.science/api/pith-number/WDFVIT2MXEEX7G2ZJ2MWIDGAQQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WDFVIT2MXEEX7G2ZJ2MWIDGAQQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WDFVIT2MXEEX7G2ZJ2MWIDGAQQ/action/storage_attestation","attest_author":"https://pith.science/pith/WDFVIT2MXEEX7G2ZJ2MWIDGAQQ/action/author_attestation","sign_citation":"https://pith.science/pith/WDFVIT2MXEEX7G2ZJ2MWIDGAQQ/action/citation_signature","submit_replication":"https://pith.science/pith/WDFVIT2MXEEX7G2ZJ2MWIDGAQQ/action/replication_record"}},"created_at":"2026-05-20T00:01:22.343309+00:00","updated_at":"2026-05-20T00:01:22.343309+00:00"}