Fast Non-Log-Concave Sampling under Nonconvex Equality and Inequality Constraints with Landing

Kijung Jeon, Michael Muehlebach, Molei Tao · 2025

Read Fast Non-Log-Concave Sampling under Nonconvex Equality and Inequality Constraints with Landing by Kijung Jeon, Michael Muehlebach, Molei Tao (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2510-22044.txt.

Introduction

Kijung Jeon, Michael Muehlebach, and Molei Tao tackle sampling under constraints. The task is to draw fair samples from a distribution fenced in by equations and inequalities. Such fences make the allowed region nonconvex. Most samplers pay for costly projection steps to stay legal. Each older method handles only one kind of fence. Firm convergence guarantees are often missing too.

Their framework, OLLA, reshapes the random walk itself. Trajectories are corrected along the direction normal to the boundary. This landing move keeps samples feasible without projections. Both equalities and inequalities fit in one method. Cases that once needed separate tools are covered together. The authors aim for speed plus firm guarantees. Bayesian statistics, chemistry, and physics all need such sampling.

Catalog record

CREATORKijung Jeon, Michael Muehlebach, Molei Tao
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE206 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2510-22044

Browse the full library · About the Tower of Babel