Research map / Optimization / Optimizers
Randomized gradient estimation
Reduces variance in gradient estimators via randomized differentiation, evolution strategies, and weight averaging.
77 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Optimizers, keyword cluster “differentiable · functions · loss”
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.
27 papers in 2023–24 and 9 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 11.2% to 3.1%.
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 Multiple Choice Learning for Training Diverse Deep Ensembles (NeurIPS 2016 · 30 citations)
- 2017 · Adaptive Relaxed ADMM: Convergence Theory and Practical Implementation (CVPR 2017 · 53 citations)
- 2018 · Total stochastic gradient algorithms and applications in reinforcement learning (NeurIPS 2018 · 7 citations)
- 2019 · Differentiable Linearized ADMM (ICML 2019 · 25 citations)
- 2020 · Stochastic Optimization for Performative Prediction (NeurIPS 2020 · 25 citations)
- 2021 · Differentiable Quality Diversity (NeurIPS 2021 · 29 citations)
- 2022 · End-to-end Stochastic Optimization with Energy-based Model (NeurIPS 2022 · 7 citations)
- 2023 · Safeguarded Learned Convex Optimization (AAAI 2023 · 7 citations)
- 2024 · GradTree: Learning Axis-Aligned Decision Trees with Gradient Descent (AAAI 2024 · 6 citations)
- 2025 · Locally Convex Global Loss Network for Decision-Focused Learning (AAAI 2025 · 6 citations)
- 2026 · Jacobian Descent for Multi-Objective Optimization (NeurIPS 2026 · 1 citation)
Earliest papers here
- Stochastic Multiple Choice Learning for Training Diverse Deep Ensembles (NeurIPS 2016 · 30 citations)
- Preconditioning Kernel Matrices (ICML 2016 · 20 citations)
- Dual Space Gradient Descent for Online Learning (NeurIPS 2016 · 15 citations)
What has been tried
69 papers here report 636 measured changes (a component added, removed or swapped, compared with a baseline): helped 64%, no clear effect 15%, hurt 17%, mixed 4%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| increase the sample count from 2 to 32 | Inventory Stock · normalized test regret | 0.237 ± 0.005 -> 0.174 ± 0.002 | Helped | Locally Convex Global Loss Network for Decision-Focused Learning (AAAI 2025 · 6 citations) |
| Replace ST Entmax with ST Gumbel Softmax | Binary classification benchmarks, optimized parameters · macro F1 | 0.771 ± .036 -> 0.569 ± .094 | Hurt | GradTree: Learning Axis-Aligned Decision Trees with Gradient Descent (AAAI 2024 · 6 citations) |
| Use gradient-based QQA transitions instead of SA with Gibbs sampling | Largest ER graphs · Sampling steps to reach solution quality | over 10^4 times | Helped | Optimization by Parallel Quasi-Quantum Annealing with Gradient-Based Sampling (ICLR 2025 · 1 citation) |
| Vary Tikhonov regularization strength λ over the tested range | MNIST sorting n=5 · element-wise ranking accuracy | No clear effect | Newton Losses: Using Curvature Information for Learning with Differentiable Algorithms (NeurIPS 2024 · 1 citation) | |
| Replace IPW with energy balancing weights | Randomized trials and observational studies in Figure 1 · accuracy | Helped | An effective framework for estimating individualized treatment rules (NeurIPS 2024 · 0 citations) |
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 36 papers; from OpenAlex and the papers' affiliations).
- Google · 3 papers
- Alibaba Group · 2 papers
- Mohamed bin Zayed University of Artificial Intelligence · 2 papers
- Georgia Institute of Technology · 2 papers
- Harvard University · 2 papers
- Korea Advanced Institute of Science and Technology · 2 papers
- Carnegie Mellon University · 2 papers
- Inria · 2 papers
- Stanford University · 2 papers
Use it
Download BibTeX of 40 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
- Differentiating through embedded optimization (117 papers)
- Surrogate losses and robust solvers (119 papers)
- Constrained and non-backprop training (97 papers)
- Multi-objective gradient methods (28 papers)
Elsewhere in the topic
- Stochastic gradient MCMC (142 papers)
- Memory-efficient LLM optimizers (79 papers)
- Adam convergence and adaptive learning rates (78 papers)
- Natural gradient and Fisher approximations (77 papers)
- SGD momentum and step-size schedules (72 papers)
- Second-order and structured preconditioning (70 papers)
- Stein variational gradient descent (47 papers)
- Gradient-based meta-learning theory (40 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).