Neural Poisson Example
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Empirical performance of two theoretical proxies for SNG relative to a realistic algorithm using a single, uniformly sampled mini-batch. The results are for solving a small, discrete Poisson problem with a small neural network ansatz. As the step size or regularization is varied the behavior of SVS-SNG matches closely with the realistic algorithm, while the behavior of the two mini-batch proxy is entirely distinct. The dimensions are m=100, n=7,801, and k=10, which is small enough to directly implement SVS. The y-axis represents the L2 error achieved after T=1,000 iterations, with the curves representing the mean result over five runs and the shaded regions spanning the minimal and maximal results attained over the same five runs.
Fields of Study & Subjects
Stochastic Gradient Optimization Techniques
Model Reduction and Neural Networks
Neural Networks and Applications
Publication details
- Journal / source
- Figshare
- Publication date
- 30 March 2026
- Language
- EN