Research map / Optimization / Nonconvex and smooth optimization
Stochastic variance reduction
Introduces SVRG-style variance-reduced gradient methods with improved rates for nonconvex optimization.
86 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Nonconvex and smooth optimization, keyword cluster “variance · reduction · reduced”
State History Tried Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
9 papers in 2023–24 and 6 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 2.8% to 1.7%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2016 · Stochastic Variance Reduction for Nonconvex Optimization (ICML 2016 · 226 citations)
- 2017 · Variance-based Regularization with Convex Objectives (NeurIPS 2017 · 104 citations)
- 2018 · Zeroth-Order Stochastic Variance Reduction for Nonconvex Optimization (NeurIPS 2018 · 96 citations)
- 2019 · SpiderBoost and Momentum: Faster Variance Reduction Algorithms (NeurIPS 2019 · 59 citations)
- 2020 · Random Reshuffling: Simple Analysis with Vast Improvements (NeurIPS 2020 · 27 citations)
- 2021 · Fast Stochastic Bregman Gradient Methods: Sharp Analysis and Variance Reduction (ICML 2021 · 5 citations)
- 2022 · A framework for bilevel optimization that enables stochastic and global variance reduction algorithms (NeurIPS 2022 · 11 citations)
- 2023 · Adaptive SGD with Polyak stepsize and Line-search: Robust Convergence and Variance Reduction (NeurIPS 2023 · 3 citations)
- 2024 · Faster Stochastic Variance Reduction Methods for Compositional MiniMax Optimization (AAAI 2024 · 3 citations)
- 2025 · A Coefficient Makes SVRG Effective (ICLR 2025 · 0 citations)
- 2026 · A Short and Unified Convergence Analysis of the SAG, SAGA, and IAG Algorithms (ICML 2026 · 0 citations)
Earliest papers here
- Stochastic Variance Reduction for Nonconvex Optimization (ICML 2016 · 226 citations)
- Riemannian SVRG: Fast Stochastic Optimization on Riemannian Manifolds (NeurIPS 2016 · 80 citations)
- Stochastic Variance Reduced Optimization for Nonconvex Sparse Learning (ICML 2016 · 33 citations)
What has been tried
84 papers here report 605 measured changes (a component added, removed or swapped, compared with a baseline): helped 69%, no clear effect 12%, hurt 12%, mixed 8%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Vary parameter m | CIFAR100, 10% imbalance · AUC | No clear effect | Faster Stochastic Variance Reduction Methods for Compositional MiniMax Optimization (AAAI 2024 · 3 citations) | |
| set batch size to n/d rather than smaller or tuned batch sizes | regression datasets · wall-clock time to reduce primal suboptimality | Helped | Drago: Primal-Dual Coupled Variance Reduction for Faster Distributionally Robust Optimization (NeurIPS 2024 · 1 citation) | |
| Vary initial coefficient α₀ across the reported values | small image classification datasets · training loss | Mixed | A Coefficient Makes SVRG Effective (ICLR 2025 · 0 citations) |
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 15 papers; from OpenAlex and the papers' affiliations).
- CISPA Helmholtz Center for Information Security · 3 papers
- CISPA · 2 papers
Use it
Download BibTeX of 33 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Proximal finite-sum methods (47 papers)
- Variance-reduced policy gradient (17 papers)
Elsewhere in the topic
- Adaptive first-order convex methods (190 papers)
- Coordinate and Adam convergence analysis (188 papers)
- Stochastic gradient last-iterate analysis (178 papers)
- Nonconvex minimax and constrained methods (100 papers)
- Stochastic convex optimization complexity (94 papers)
- Acceleration and Nesterov dynamics (91 papers)
- Equilibrium learning in zero-sum games (88 papers)
- Complexity of nonconvex first-order methods (79 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).