Research map / Optimization / Nonconvex and smooth optimization
Online Learning Under a Separable Stochastic Approximation Framework
Min Gan, Xiang-Xiang Su, Guangyong Chen, Jing Chen, C. L. Philip Chen
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 SepSA instead of Adam | FNN · Energy efficiency (online) | test MSE: 122.6856 -> 8.3049 | Improved |
| use SepSA instead of Adam | FNN · Diabetes (online) | test MSE: 27033.5684 -> 2940.2236 | Improved |
| use SepSA instead of Adam | CNN · MNIST (online) | test accuracy: 97.97% -> 98.39% | Improved |
| use SepSA instead of Adam | CNN · CIFAR-10 (online) | test accuracy: 24.61% -> 38.14% | Improved |
| use SepSA instead of slimTrain | CNN · MNIST (online) | test accuracy: 14.74% -> 98.39% | Improved |
| use SepSA instead of Adam | FNN · Energy efficiency (mini-batch) | test MSE: 18.3703 -> 3.2776 | Improved |
| use SepSA instead of Adam | CNN · CIFAR-10 (mini-batch) | test accuracy: 44.79% -> 74.95% | Improved |
| use SepSA instead of slimTrain | CNN · CIFAR-10 (mini-batch) | test accuracy: 69.86% -> 74.95% | Improved |
| use SepSA instead of Adam | CNN · MNIST (mini-batch) | test accuracy: 99.15% -> 99.23% | Improved |
| use SepSA instead of Adam | FNN · Energy efficiency (online) | running time (s): 0.3395 -> 0.3987 | Hurt |
| use SepSA instead of Adam | FNN · Diabetes (mini-batch) | test MSE: 17581.1328 -> 3076.0276 | Improved |
| use SepSA instead of Adam | FNN · Energy efficiency, wide learning-rate range | sensitivity to learning rate | Improved |
Closest papers by meaning
- HiGrad: Uncertainty Quantification for Online Learning and Stochastic Approximation (JMLR 2023 · similarity 0.88)
- Provable and Practical Online Learning Rate Adaptation with Hypergradient Descent (ICML 2025 · similarity 0.87)
- Asymptotic theory of SGD with a general learning-rate (NeurIPS 2025 · similarity 0.87)
- Stochastic Approximate Gradient Descent via the Langevin Algorithm (AAAI 2020 · similarity 0.87)
- On the Convergence of Loss and Uncertainty-based Active Learning Algorithms (NeurIPS 2024 · similarity 0.87)
- Limited Memory Online Gradient Descent for Kernelized Pairwise Learning with Dynamic Averaging (AAAI 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.
