The Performance Of The Unadjusted Langevin Algorithm Without Smoothness Assumptions
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909935541944320 |
|---|---|
| author | Johnston, Tim Lytras, Iosif Makras, Nikolaos Sabanis, Sotirios |
| author_facet | Johnston, Tim Lytras, Iosif Makras, Nikolaos Sabanis, Sotirios |
| contents | In this article, we study the problem of sampling from distributions whose densities are not necessarily smooth nor logconcave. We propose a simple Langevin-based algorithm that does not rely on popular but computationally challenging techniques, such as the Moreau-Yosida envelope or Gaussian smoothing, and show consequently that the performance of samplers like ULA does not necessarily degenerate arbitrarily with low regularity. In particular, we show that the Lipschitz or Hölder continuity assumption can be replaced by a geometric one-sided Lipschitz condition that allows even for discontinuous log-gradients. We derive non-asymptotic guarantees for the convergence of the algorithm to the target distribution in Wasserstein distances. Non-asymptotic bounds are also provided for the performance of the algorithm as an optimizer, specifically for the solution of associated excess risk optimization problems. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_03458 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | The Performance Of The Unadjusted Langevin Algorithm Without Smoothness Assumptions Johnston, Tim Lytras, Iosif Makras, Nikolaos Sabanis, Sotirios Machine Learning Optimization and Control Probability Computation In this article, we study the problem of sampling from distributions whose densities are not necessarily smooth nor logconcave. We propose a simple Langevin-based algorithm that does not rely on popular but computationally challenging techniques, such as the Moreau-Yosida envelope or Gaussian smoothing, and show consequently that the performance of samplers like ULA does not necessarily degenerate arbitrarily with low regularity. In particular, we show that the Lipschitz or Hölder continuity assumption can be replaced by a geometric one-sided Lipschitz condition that allows even for discontinuous log-gradients. We derive non-asymptotic guarantees for the convergence of the algorithm to the target distribution in Wasserstein distances. Non-asymptotic bounds are also provided for the performance of the algorithm as an optimizer, specifically for the solution of associated excess risk optimization problems. |
| title | The Performance Of The Unadjusted Langevin Algorithm Without Smoothness Assumptions |
| topic | Machine Learning Optimization and Control Probability Computation |
| url | https://arxiv.org/abs/2502.03458 |