Research map / Optimization / Nonconvex and smooth optimization
SPAN: A Stochastic Projected Approximate Newton Method
Xunpeng Huang, Xianfeng Liang, Zhengyang Liu, Lei Li, Yue Yu, Yitan Li
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 SPAN's stochastic projected approximate Hessian instead of the compared optimizers | Binary image classification · MNIST4-9 and CovType | wall-clock time to convergence | Improved |
| Use SPAN instead of LiSSA | Binary image classification · MNIST4-9 | empirical convergence rate | Improved |
| Use SPAN instead of NewSamp | Binary image classification · MNIST4-9 and CovType | empirical convergence rate | Hurt |
| Use SPAN's stochastic projected Hessian approximation instead of the compared second-order methods | Binary image classification · MNIST4-9 and CovType | Hessian approximation error | Improved |
| Use SPAN instead of NewSamp | Binary image classification · MNIST4-9 | wall-clock time to convergence | Improved |
| Use SPAN instead of LiSSA | Binary image classification · CovType | wall-clock time to convergence | Improved |
Closest papers by meaning
- Approximate Newton Methods (JMLR 2021 · similarity 0.89)
- Approximate Newton Methods and Their Local Convergence (ICML 2017 · similarity 0.88)
- Explicit Convergence Rates of Greedy and Random Quasi-Newton Methods (JMLR 2022 · similarity 0.88)
- Stochastic Adaptive Quasi-Newton Methods for Minimizing Expected Values (ICML 2017 · similarity 0.88)
- Skip the Hessian, Keep the Rates: Globalized Semismooth Newton with Lazy Hessian Updates (NeurIPS 2026 · similarity 0.88)
- Stochastic Newton Proximal Extragradient Method (NeurIPS 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.
