{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:OP6HBND7GN2D7JZCIQSVHDJMG4","short_pith_number":"pith:OP6HBND7","schema_version":"1.0","canonical_sha256":"73fc70b47f33743fa7224425538d2c3723143591e00803f7ad0d891e292b1243","source":{"kind":"arxiv","id":"2605.17306","version":1},"attestation_state":"computed","paper":{"title":"Speeding Up Nonsmooth Bayesian MCMC Sampling via Inexact Proximal Unadjusted Langevin Algorithm","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"An inexact proximal unadjusted Langevin algorithm samples from nonsmooth composite posteriors by using controlled approximations in place of exact proximal operators.","cross_cats":[],"primary_cat":"math.OC","authors_text":"Alireza Kabgani, Masoud Ahookhosh, Susan Ghaderi, Yves Moreau","submitted_at":"2026-05-17T07:48:28Z","abstract_excerpt":"We study sampling from posterior distributions with nonsmooth composite potentials, a setting in which proximal-based Langevin methods are theoretically appealing but in practice limited to simple functions with closed-form proximal operators. We introduce iPULA for composite potentials, an inexact proximal unadjusted Langevin algorithm that replaces exact proximal steps with controlled approximations. Our approach leverages the Moreau envelope to smooth the potential, while allowing inexact evaluation of its gradient through inexact proximal computations. We establish non-asymptotic convergen"},"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":true},"canonical_record":{"source":{"id":"2605.17306","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"math.OC","submitted_at":"2026-05-17T07:48:28Z","cross_cats_sorted":[],"title_canon_sha256":"ee6e1e8ae7f632e3f6dfaaa8b6212f872cd10a871694b7e8e855c021ba749395","abstract_canon_sha256":"75a6606b579c72620a939ba1808777ce3296794cc194e8ce664f09a5723bbca8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:03:51.316336Z","signature_b64":"Z5q2VrwuC2eqVyWpcxd4gnkJWJWHgR4MNo+lTlNY14sam9bV/UCl/cbJ6PoYwxJQ2T7ZmF6WDY09RfC1NSnVAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"73fc70b47f33743fa7224425538d2c3723143591e00803f7ad0d891e292b1243","last_reissued_at":"2026-05-20T00:03:51.315550Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:03:51.315550Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Speeding Up Nonsmooth Bayesian MCMC Sampling via Inexact Proximal Unadjusted Langevin Algorithm","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"An inexact proximal unadjusted Langevin algorithm samples from nonsmooth composite posteriors by using controlled approximations in place of exact proximal operators.","cross_cats":[],"primary_cat":"math.OC","authors_text":"Alireza Kabgani, Masoud Ahookhosh, Susan Ghaderi, Yves Moreau","submitted_at":"2026-05-17T07:48:28Z","abstract_excerpt":"We study sampling from posterior distributions with nonsmooth composite potentials, a setting in which proximal-based Langevin methods are theoretically appealing but in practice limited to simple functions with closed-form proximal operators. We introduce iPULA for composite potentials, an inexact proximal unadjusted Langevin algorithm that replaces exact proximal steps with controlled approximations. Our approach leverages the Moreau envelope to smooth the potential, while allowing inexact evaluation of its gradient through inexact proximal computations. We establish non-asymptotic convergen"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We establish non-asymptotic convergence guarantees for iPULA, explicitly characterizing the impact of inexactness on the sampling error and showing that the inexactness preserves convergence rates up to a quantifiable bias.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The analysis assumes that the inexact proximal computations can be performed with sufficient accuracy control so that the bias remains quantifiable and does not destroy the convergence rate; this enters when the Moreau envelope smoothing is combined with the inexact gradient evaluations.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"iPULA replaces exact proximal steps with inexact approximations in unadjusted Langevin sampling and proves non-asymptotic convergence that holds up to a quantifiable bias from the inexactness.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"An inexact proximal unadjusted Langevin algorithm samples from nonsmooth composite posteriors by using controlled approximations in place of exact proximal operators.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"30b9e45ff5167e1318df99af643beb567f1411d867f1044fef76a0ff1e28b2b1"},"source":{"id":"2605.17306","kind":"arxiv","version":1},"verdict":{"id":"2a25fa0b-2b8c-4b22-8f5e-75ad1ee2cc39","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T22:48:11.619405Z","strongest_claim":"We establish non-asymptotic convergence guarantees for iPULA, explicitly characterizing the impact of inexactness on the sampling error and showing that the inexactness preserves convergence rates up to a quantifiable bias.","one_line_summary":"iPULA replaces exact proximal steps with inexact approximations in unadjusted Langevin sampling and proves non-asymptotic convergence that holds up to a quantifiable bias from the inexactness.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The analysis assumes that the inexact proximal computations can be performed with sufficient accuracy control so that the bias remains quantifiable and does not destroy the convergence rate; this enters when the Moreau envelope smoothing is combined with the inexact gradient evaluations.","pith_extraction_headline":"An inexact proximal unadjusted Langevin algorithm samples from nonsmooth composite posteriors by using controlled approximations in place of exact proximal operators."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.17306/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_compliance","ran_at":"2026-05-19T23:01:52.480153Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T23:01:19.689019Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T22:01:57.795628Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T21:33:23.757287Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"653745ca638cea4757c761a56617df3190a23dbfae4543375de0a2a8bded9835"},"references":{"count":35,"sample":[{"doi":"","year":2021,"title":"M. Ahookhosh, A. Themelis, and P. Patrinos. A Bregman forward-backward linesearch algo- rithm for nonconvex composite optimization: Superlinear convergence to nonisolated local min- ima.SIAM Journal o","work_id":"043309ed-55ad-48fc-8c08-234598ccf313","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2017,"title":"H. H. Bauschke and P. L. Combettes.Convex Analysis and Monotone Operator Theory in Hilbert Spaces. Springer, 2nd edition, 2017","work_id":"177778f4-f4e0-4182-bb89-38a8ec69ea72","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2011,"title":"S. Brooks, A. Gelman, G. Jones, and X.-L. Meng, editors.Handbook of Markov Chain Monte Carlo. Chapman and Hall/CRC, 2011","work_id":"03389b4e-587e-4a1a-8667-2d71b799c639","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2020,"title":"N. S. Chatterji, J. Diakonikolas, M. I. Jordan, and P. L. Bartlett. Langevin Monte Carlo without smoothness. InProceedings of the Twenty Third International Conference on Artificial Intelligence and S","work_id":"04d734de-383a-4c21-8998-d2b61306d14c","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1907,"title":"F. R. Crucinio, A. Durmus, P. Jim´ enez, and G. O. Roberts. Optimal scaling results for Moreau- Yosida Metropolis-adjusted Langevin algorithms.Bernoulli, 31(3):1889–1907, 2025","work_id":"730b2cd0-3485-4355-b0fc-e44878860d0e","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":35,"snapshot_sha256":"79e69512c77e8de53249991cb98e52280fb3ca0be203396c51a313bcc1d0f855","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"78aa26a48462ba9a8abf35ec8490974ff9d24f2901e01ff8a61e61dabe52bc5e"},"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":"2605.17306","created_at":"2026-05-20T00:03:51.315674+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.17306v1","created_at":"2026-05-20T00:03:51.315674+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.17306","created_at":"2026-05-20T00:03:51.315674+00:00"},{"alias_kind":"pith_short_12","alias_value":"OP6HBND7GN2D","created_at":"2026-05-20T00:03:51.315674+00:00"},{"alias_kind":"pith_short_16","alias_value":"OP6HBND7GN2D7JZC","created_at":"2026-05-20T00:03:51.315674+00:00"},{"alias_kind":"pith_short_8","alias_value":"OP6HBND7","created_at":"2026-05-20T00:03:51.315674+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OP6HBND7GN2D7JZCIQSVHDJMG4","json":"https://pith.science/pith/OP6HBND7GN2D7JZCIQSVHDJMG4.json","graph_json":"https://pith.science/api/pith-number/OP6HBND7GN2D7JZCIQSVHDJMG4/graph.json","events_json":"https://pith.science/api/pith-number/OP6HBND7GN2D7JZCIQSVHDJMG4/events.json","paper":"https://pith.science/paper/OP6HBND7"},"agent_actions":{"view_html":"https://pith.science/pith/OP6HBND7GN2D7JZCIQSVHDJMG4","download_json":"https://pith.science/pith/OP6HBND7GN2D7JZCIQSVHDJMG4.json","view_paper":"https://pith.science/paper/OP6HBND7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.17306&json=true","fetch_graph":"https://pith.science/api/pith-number/OP6HBND7GN2D7JZCIQSVHDJMG4/graph.json","fetch_events":"https://pith.science/api/pith-number/OP6HBND7GN2D7JZCIQSVHDJMG4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OP6HBND7GN2D7JZCIQSVHDJMG4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OP6HBND7GN2D7JZCIQSVHDJMG4/action/storage_attestation","attest_author":"https://pith.science/pith/OP6HBND7GN2D7JZCIQSVHDJMG4/action/author_attestation","sign_citation":"https://pith.science/pith/OP6HBND7GN2D7JZCIQSVHDJMG4/action/citation_signature","submit_replication":"https://pith.science/pith/OP6HBND7GN2D7JZCIQSVHDJMG4/action/replication_record"}},"created_at":"2026-05-20T00:03:51.315674+00:00","updated_at":"2026-05-20T00:03:51.315674+00:00"}