Research map / Optimization / Nonconvex and smooth optimization
SGD with Coordinate Sampling: Theory and Practice
Rémi Leluc, François Portier
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 MUSKETEER adaptive coordinate sampling (average or absolute gains) instead of uniform sampling | zeroth-order SGD on ridge and logistic regression · Synthetic block-structured data | optimality gap | Improved |
| use uniform coordinate sampling instead of the full finite-difference gradient | zeroth-order SGD · Synthetic block-structured data | optimality gap | No clear effect |
| increase block structure strength alpha (ablation) | MUSKETEER · Synthetic ridge and logistic | relative performance | Improved |
| use MUSKETEER instead of standard zeroth-order methods | linear-layer neural network · MNIST and Fashion-MNIST | test accuracy | Improved |
Closest papers by meaning
- Adaptive Sampling Probabilities for Non-Smooth Optimization (ICML 2017 · similarity 0.91)
- Even Faster Accelerated Coordinate Descent Using Non-Uniform Sampling (ICML 2016 · similarity 0.90)
- Sampling from Gaussian Process Posteriors using Stochastic Gradient Descent (NeurIPS 2023 · similarity 0.89)
- Adaptive Sampling for SGD by Exploiting Side Information (ICML 2016 · similarity 0.89)
- The Sample Complexity of Gradient Descent in Stochastic Convex Optimization (NeurIPS 2024 · similarity 0.88)
- Derivatives of Stochastic Gradient Descent in parametric optimization (NeurIPS 2024 · 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.
