Research map / Optimization / Nonconvex and smooth optimization
When is the Convergence Time of Langevin Algorithms Dimension Independent? A Composite Optimization Viewpoint
Yoav S Freund, Yi-An Ma, Tong Zhang
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 |
|---|---|---|---|
| Use explicitly integrable strongly convex prior diffusion alongside likelihood gradient steps | Langevin Algorithm with Prior Diffusion · Lipschitz convex posterior | gradient queries: T = \frac{5 G^2}{\beta m \epsilon} | Improved |
| Use streaming stochastic likelihood gradients with prior diffusion | Streaming SGLD with Prior Diffusion · Lipschitz convex posterior | gradient queries: T = \frac{5 G_\ell^2}{\beta m \epsilon} | Improved |
| Use a quadratic normal prior and bound smoothness through the Hessian | Langevin Algorithm with Prior Diffusion · Smooth convex posterior | gradient queries: T = 64 \cdot \max\lrbb{ \frac{8\trace\lrp{H^2}}{m^2 \epsilon}, \frac{2 U(0) }{\epsilon} } | Improved |
| Apply the smooth-case method to ridge-separable likelihoods | Langevin Algorithm with Prior Diffusion · Ridge-separable smooth convex posterior | gradient queries: T = 64 \cdot \max\lrbb{ \frac{ 8 L_s^2 R_z^2}{m^2 \epsilon}, \frac{2 U(0)}{\epsilon} } | Improved |
Closest papers by meaning
- Global Convergence of Langevin Dynamics Based Algorithms for Nonconvex Optimization (NeurIPS 2018 · similarity 0.88)
- Breaking Reversibility Accelerates Langevin Dynamics for Non-Convex Optimization (NeurIPS 2020 · similarity 0.88)
- Improved Convergence Rate of Stochastic Gradient Langevin Dynamics with Variance Reduction and its Application to Optimization (NeurIPS 2022 · similarity 0.88)
- Mixing Times and Privacy Analysis for the Projected Langevin Algorithm under a Modulus of Continuity (JMLR 2025 · similarity 0.87)
- Time-Independent Information-Theoretic Generalization Bounds for SGLD (NeurIPS 2023 · similarity 0.87)
- Generalization of noisy SGD in unbounded non-convex settings (ICML 2025 · similarity 0.86)
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.
