Saved in:
Bibliographic Details
Main Authors: Gissler, Armand, Saremi, Saeed, Bach, Francis
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2602.15587
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917277987766272
author Gissler, Armand
Saremi, Saeed
Bach, Francis
author_facet Gissler, Armand
Saremi, Saeed
Bach, Francis
contents Sampling from discrete distributions is a ubiquitous task in machine learning, recently revisited by the emergence of discrete diffusion models. While Langevin algorithms constitute the state of the art for continuous spaces, discrete versions lack similar theoretical guarantees when the step-size becomes small. In this paper, we address this limitation by interpreting discrete sampling algorithms as discretizations of continuous-time dynamics on the hypercube. In particular, we describe several score functions for discrete algorithms which result in approximations of Glauber dynamics for the correct target distribution. We also compute upper bounds for the contraction of these algorithms, with or without Metropolis adjustment.
format Preprint
id arxiv_https___arxiv_org_abs_2602_15587
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Adjusted Scores for Discrete Langevin Algorithms
Gissler, Armand
Saremi, Saeed
Bach, Francis
Statistics Theory
Sampling from discrete distributions is a ubiquitous task in machine learning, recently revisited by the emergence of discrete diffusion models. While Langevin algorithms constitute the state of the art for continuous spaces, discrete versions lack similar theoretical guarantees when the step-size becomes small. In this paper, we address this limitation by interpreting discrete sampling algorithms as discretizations of continuous-time dynamics on the hypercube. In particular, we describe several score functions for discrete algorithms which result in approximations of Glauber dynamics for the correct target distribution. We also compute upper bounds for the contraction of these algorithms, with or without Metropolis adjustment.
title Adjusted Scores for Discrete Langevin Algorithms
topic Statistics Theory
url https://arxiv.org/abs/2602.15587