Summary of Revisiting Random Weight Perturbation For Efficiently Improving Generalization, by Tao Li et al.
Revisiting Random Weight Perturbation for Efficiently Improving Generalization
by Tao Li, Qinghua Tao, Weihao Yan, Zehao Lei, Yingwen Wu, Kun Fang, Mingzhen He, Xiaolin Huang
First submitted to arxiv on: 30 Mar 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary In this paper, the authors revisit a method called Random Weight Perturbation (RWP) to improve the generalization ability of deep neural networks. RWP is related to Sharpness-Aware Minimization (SAM), which uses adversarial weight perturbation to minimize the worst-case neighborhood loss. While SAM has been shown to be effective in improving generalization, RWP’s empirical performance has lagged behind. The authors propose improvements to RWP from two perspectives: the trade-off between generalization and convergence, and the generation of random perturbations. Through extensive experiments, they show that their enhanced RWP methods achieve greater efficiency in enhancing generalization, particularly in large-scale problems, while offering comparable or even superior performance to SAM. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about making deep neural networks better at learning new things. Right now, there are two ways to make them do this: one uses a technique called sharpness-aware minimization (SAM), and the other uses random weight perturbation (RWP). RWP is related to SAM but doesn’t work as well. The authors of this paper want to make RWP better, so they try out some new ideas. They test these ideas on lots of different problems and show that their improved RWP methods can do a good job at making the networks learn. |
Keywords
* Artificial intelligence * Generalization * Sam
