Research map / Optimization / Nonconvex and smooth optimization
Variance reduction for Random Coordinate Descent-Langevin Monte Carlo
ZHIYAN DING, Qin Li
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| replace full gradient with vanilla random coordinate descent (one random partial derivative) in U-LMC | U-LMC · Gaussian counter-example, d>2 | number of finite-difference evaluations to reach epsilon accuracy | No clear effect |
| add randomized coordinate averaging variance reduction (RCAD) to RCD for overdamped LMC | O-LMC · log-concave sampling theory (Lipschitz hessian) | number of partial derivative evaluations: O~(d^2/epsilon) -> O~(d^{3/2}/epsilon) | Improved |
| add RCAD variance reduction to RCD for underdamped LMC | U-LMC · log-concave sampling theory | number of partial derivative evaluations: O~(d^{3/2}/epsilon) -> O~(max{d^{4/3}/epsilon^{2/3}, d^{1/2}/epsilon}) | Improved |
| add RCAD to RCD for overdamped LMC while dropping the Lipschitz hessian assumption | O-LMC · log-concave sampling theory without hessian Lipschitz | number of partial derivative evaluations: O~(d^2/epsilon^2) -> O~(max{d^{3/2}/epsilon, d/epsilon^2}) | Improved |
| add RCAD variance reduction to RCD-O-LMC (ablation) | RCD-O-LMC · Example 1: N(0,I_d), d=1000, MSE of phi(x)=|x_1|^2 | error vs stepsize | Improved |
| add RCAD variance reduction to RCD-U-LMC (ablation) | RCD-U-LMC · Example 2: bimodal Gaussian mixture, MSE of phi(x)=|x_1|^2 | error vs stepsize | Improved |
Closest papers by meaning
- Langevin Monte Carlo: random coordinate descent and variance reduction (JMLR 2021 · similarity 0.95)
- Langevin Quasi-Monte Carlo (NeurIPS 2023 · similarity 0.91)
- Penalized Overdamped and Underdamped Langevin Monte Carlo Algorithms for Constrained Sampling (JMLR 2024 · similarity 0.88)
- Constrained Sampling with Primal-Dual Langevin Monte Carlo (NeurIPS 2024 · similarity 0.87)
- Non-asymptotic Error Bounds in $\mathcal{W}_2$-Distance with Sqrt(d) Dimension Dependence and First Order Convergence for Langevin Monte Carlo beyond Log-Concavity (ICML 2025 · similarity 0.87)
- Provable Benefit of Annealed Langevin Monte Carlo for Non-log-concave Sampling (ICLR 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
