In-and-Out: Algorithmic Diffusion for Sampling Convex Bodies

Authors: Yunbum Kook, Santosh Vempala, Matthew Zhang

NeurIPS 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical The contribution of this paper is primarily theoretical, and the paper does not contain any experimental results. We present a new random walk for uniformly sampling high-dimensional convex bodies. It achieves state-of-the-art runtime complexity with stronger guarantees on the output than previously known, namely in Rényi divergence (which implies TV, W2, KL, χ2). The proof departs from known approaches for polytime algorithms for the problem we utilize a stochastic diffusion perspective to show contraction to the target distribution with the rate of convergence determined by functional isoperimetric constants of the stationary density.
Researcher Affiliation Academia Yunbum Kook School of Computer Science Georgia Institute of Technology yb.kook@gatech.edu Santosh S. Vempala School of Computer Science Georgia Institute of Technology vempala@gatech.edu Matthew S. Zhang Department of Computer Science, University of Toronto, matthew.zhang@mail.utoronto.ca
Pseudocode Yes Algorithm 1 In-and-Out
Open Source Code No The paper does not contain experiments and therefore does not require data or code.
Open Datasets No The paper does not contain experiments.
Dataset Splits No The paper does not contain experiments.
Hardware Specification No The paper does not contain experiments.
Software Dependencies No The paper does not contain experiments.
Experiment Setup No The paper does not contain experiments.