Research map / Optimization / Nonconvex and smooth optimization
The Physical Systems Behind Optimization Algorithms
Lin Yang, Raman Arora, Vladimir braverman, Tuo Zhao
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 Nesterov momentum to VGD | VGD · Convex functions satisfying the PL condition | objective convergence rate: from $L/\mu$ to $\sqrt{L/\mu}$ | Improved |
| Use the critical-damping momentum setting, $s=1/2$ | NAG · Convex functions satisfying the PL condition | convergence rate: s=1/2 | Improved |
| Use a massless gradient-descent system instead of a massive accelerated system | VGD and NAG · Convex functions satisfying the PL condition | convergence rate: from $\sqrt{L/\mu}$ to $L/\mu$ | Hurt |
| Use randomized coordinate updates rather than full-gradient updates | RCGD · General convex functions | objective convergence rate: \frac{dL_{\max}\norm{x_0}^2}{2k} | Mixed |
| Add momentum to randomized coordinate gradient descent | ARCG · General convex functions | objective convergence rate: \frac{C_2 d}{k^2} | Improved |
| Replace a full-gradient update with a randomized coordinate update | RCGD · Polyak–Łojasiewicz condition | objective convergence rate: \exp\big(-\frac{2\mu}{dL_{\max}}k\big) | Hurt |
Closest papers by meaning
- On Constraints in First-Order Optimization: A View from Non-Smooth Dynamical Systems (JMLR 2022 · similarity 0.88)
- Spectral Preconditioning for Gradient Methods on Graded Non-convex Functions (ICML 2024 · similarity 0.86)
- Adaptive Proximal Gradient Method for Convex Optimization (NeurIPS 2024 · similarity 0.86)
- Two-timescale Extragradient for Finding Local Minimax Points (ICLR 2024 · similarity 0.85)
- An operator preconditioning perspective on training in physics-informed machine learning (ICLR 2024 · similarity 0.85)
- Optimizing $(L_0, L_1)$-Smooth Functions by Gradient Methods (ICLR 2025 · similarity 0.85)
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.
