UnsolvedMajor Unsolved Problem

Optimal Mixing Time for Log-Concave Sampling

§ Problem Statement

Setup

Unsolved Problem

§ Discussion

Loading discussion…

§ Significance & Implications

§ Known Partial Results

§ References

[1]

Optimal convergence rate of the proximal point algorithm for strongly log-concave distributions

Yongxin Chen, Sinho Chewi, Adil Salim, Andre Wibisono (2022)

Annals of Applied Probability

📍 Section 4.2 (Applications of the convergence results), item 2 in the numbered comparison list (“it is currently not known how to perform a discretization analysis … with linear dependence on $\kappa$ under $\alpha$-LSI or $\alpha$-PI”), p. 8

[2]

Rapid Convergence of the Unadjusted Langevin Algorithm: Isoperimetry Suffices

Santosh Vempala, Andre Wibisono (2019)

[3]

Underdamped Langevin MCMC: A non-asymptotic analysis

Xiang Cheng, Niladri S. Chatterji, Peter L. Bartlett, Michael I. Jordan (2018)

[4]

A random polynomial-time algorithm for approximating the volume of convex bodies

Martin Dyer, Alan Frieze, Ravi Kannan (1991)

Journal of the ACM

📍 Section 3 (Random walks on convex bodies), Theorem 3.1 (rapid mixing of the ball walk via conductance), pp. 8–10.

[5]

Theoretical guarantees for approximate sampling from smooth and log-concave density

Arnak S. Dalalyan (2017)

Journal of the Royal Statistical Society Series B

📍 Section 3 (Main results), Theorem 1 (non-asymptotic convergence bound for unadjusted Langevin algorithm in TV distance), p. 658.

§ Tags