Research map / Optimization / Nonconvex and smooth optimization
A Simple Practical Accelerated Method for Finite Sums
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Evaluate the selected component's gradient at the updated iterate via a proximal step rather than at the current iterate | Point-SAGA compared with SAGA · Strongly convex smooth finite sums | iteration complexity: O\left(\left(L/\mu+n\right)\log\left(1/\epsilon\right)\right) -> O\left(\left(\sqrt{\frac{nL}{\mu}}+n\right)\log\left(\f | Helped |
| Use proximal updates with fixed step sizes | Point-SAGA · Strongly convex non-smooth problems | convergence rate | Helped |
| Apply acceleration | Point-SAGA versus non-accelerated methods · LIBSVM logistic-regression experiments | function value sub-optimality | Helped |
| Use Point-SAGA instead of SDCA | RCV1 with hinge loss · RCV1 | convergence speed | No clear effect |
Closest papers by meaning
- Limitations on Variance-Reduction and Acceleration Schemes for Finite Sums Optimization (NeurIPS 2017 · similarity 0.91)
- An Accelerated DFO Algorithm for Finite-sum Convex Functions (ICML 2020 · similarity 0.89)
- A unified variance-reduced accelerated gradient method for convex optimization (NeurIPS 2019 · similarity 0.88)
- Freya PAGE: First Optimal Time Complexity for Large-Scale Nonconvex Finite-Sum Optimization with Heterogeneous Asynchronous Computations (NeurIPS 2024 · similarity 0.87)
- Efficient Continual Finite-Sum Minimization (ICLR 2024 · similarity 0.86)
- Decentralized Convex Finite-Sum Optimization with Better Dependence on Condition Numbers (ICML 2024 · 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.