Research map / Optimization / Nonconvex and smooth optimization
NESTT: A Nonconvex Primal-Dual Splitting Method for Distributed and Stochastic Optimization
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 |
|---|---|---|---|
| NESTT-G instead of SGD, uniform sampling, N=10 | nonconvex regression · N=10 | optimality gap: 3.4054 -> 2.3E-21 | Helped |
| NESTT-G instead of SGD, non-uniform, N=50 | nonconvex regression · N=50 | optimality gap: 0.0409 -> 2.7E-4 | Helped |
| NESTT-G instead of SAGA, non-uniform sampling, N=50 | nonconvex regression · N=50 | optimality gap: 3.3187 -> 2.7E-4 | Helped |
| NESTT-G instead of SAGA, uniform sampling, N=10 | nonconvex regression · N=10 | optimality gap: 2.7E-17 -> 2.3E-21 | Helped |
| NESTT-G instead of NESTT-E (alpha=10), uniform, N=30 | nonconvex regression · N=30 | optimality gap: 3.2E-6 -> 4.5E-7 | Helped |
| use non-uniform instead of uniform sampling for SAGA, N=20 (ablation) | SAGA · N=20 | optimality gap: 7.7E-7 -> 11.3435 | Hurt |
| use non-uniform instead of uniform sampling for NESTT-G, N=10 (ablation) | NESTT-G · N=10 | optimality gap: 2.3E-21 -> 6.1E-24 | Helped |
| use non-uniform instead of uniform sampling for NESTT-G, N=50 (ablation) | NESTT-G · N=50 | optimality gap: 1.2E-4 -> 2.7E-4 | Hurt |
| increase number of mini-batches N from 10 to 50, NESTT-G uniform (ablation) | NESTT-G · uniform sampling | optimality gap: 2.3E-21 -> 1.2E-4 | Hurt |
| increase number of mini-batches N from 10 to 50, SGD uniform (ablation) | SGD · uniform sampling | optimality gap: 3.4054 -> 0.0154 | Helped |
| use larger alpha in NESTT-E (ablation) | NESTT-E · synthetic regression | convergence speed | Helped |
| NESTT-G with non-uniform sampling vs GD in worst case, Case IV | NESTT-G · gradient evaluations | gradient evaluations: O(N^{2}/\epsilon) -> O(N/\epsilon) | Helped |
Closest papers by meaning
- Stochastic Nested Variance Reduction for Nonconvex Optimization (NeurIPS 2018 · similarity 0.90)
- Stochastic Smoothed Primal-Dual Algorithms for Nonconvex Optimization with Linear Inequality Constraints (ICML 2025 · similarity 0.89)
- Proximal Stochastic Methods for Nonsmooth Nonconvex Finite-Sum Optimization (NeurIPS 2016 · similarity 0.88)
- Prox-PDA: The Proximal Primal-Dual Algorithm for Fast Distributed Nonconvex Optimization and Learning Over Networks (ICML 2017 · similarity 0.88)
- High-Probability Convergence for Composite and Distributed Stochastic Minimization and Variational Inequalities with Heavy-Tailed Noise (ICML 2024 · similarity 0.88)
- Accelerated Algorithms for Constrained Nonconvex-Nonconcave Min-Max Optimization and Comonotone Inclusion (ICML 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.