Research map / Optimization / Optimizers
HiGrad: Uncertainty Quantification for Online Learning and Stochastic Approximation
Weijie J. Su, Yuancheng Zhu
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 longer thread length (fourth configuration, thread length 3N/7) in HiGrad (ablation) | HiGrad · Simulated linear and logistic regression | estimation risk relative to SGD | Improved |
| use shortest thread length N/4 (first configuration) (ablation) | HiGrad · Simulated regression | estimation risk | Hurt |
| vary HiGrad configuration (number of splits and thread lengths) (ablation) | HiGrad confidence intervals · Simulated regression, three types of true coefficients | coverage probability | No clear effect |
| use two splits with T = 4 threads (last configuration) (ablation) | HiGrad · Simulated regression | accuracy, coverage, informativeness ratings | Improved |
| add HiGrad prediction intervals to SGD predictions | SGD logistic regression · Adult dataset | empirical coverage probability | Improved |
Closest papers by meaning
- Online stochastic gradient descent on non-convex losses from high-dimensional inference (JMLR 2021 · similarity 0.89)
- Online Learning Under a Separable Stochastic Approximation Framework (TPAMI 2024 · similarity 0.88)
- Asymptotic theory of SGD with a general learning-rate (NeurIPS 2025 · similarity 0.88)
- Parallelizing Stochastic Gradient Descent for Least Squares Regression: Mini-batching, Averaging, and Model Misspecification (JMLR 2018 · similarity 0.88)
- Statistical Guarantees for High-Dimensional Stochastic Gradient Descent (NeurIPS 2025 · similarity 0.87)
- On the Convergence of Loss and Uncertainty-based Active Learning Algorithms (NeurIPS 2024 · similarity 0.86)
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.
