Research map / Optimization / Nonconvex and smooth optimization
Penalized Langevin dynamics with vanishing penalty for smooth and log-concave targets
Avetik Karagulyan, Arnak Dalalyan
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 |
|---|---|---|---|
| Add a decreasing linear penalty that vanishes over time | Langevin diffusion for log-concave targets · Log-concave targets | Wasserstein-2 distance | Improved |
| Use α(t) = 1/({2t + 121μ_2(π)}) | Langevin's Adjusted Process (LAP) · Log-concave targets, m = 0 | Wasserstein-2 distance | Improved |
| Estimate the drift integral at a randomized intermediate point instead of using only the starting point | Overdamped Langevin · Strongly log-concave targets satisfying (m-M-CL) | Wasserstein-2 distance | Improved |
| Average R randomized intermediate-point drift estimates | Randomized mid-point method · Strongly log-concave targets satisfying (m-M-CL) | Wasserstein-2 distance | Improved |
Closest papers by meaning
- Analysis of Langevin Monte Carlo via Convex Optimization (JMLR 2019 · similarity 0.90)
- Bounding the Error of Discretized Langevin Algorithms for Non-Strongly Log-Concave Targets (JMLR 2022 · similarity 0.90)
- Implicit Langevin Algorithms for Sampling From Log-concave Densities (JMLR 2021 · similarity 0.90)
- Generalization of noisy SGD in unbounded non-convex settings (ICML 2025 · similarity 0.89)
- Langevin Monte Carlo for strongly log-concave distributions: Randomized midpoint revisited (ICLR 2024 · similarity 0.89)
- Provable Convergence and Limitations of Geometric Tempering for Langevin Dynamics (ICLR 2025 · similarity 0.88)
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.
