Summary of Over-parameterization and Adversarial Robustness in Neural Networks: An Overview and Empirical Analysis, by Zhang Chen et al.
Over-parameterization and Adversarial Robustness in Neural Networks: An Overview and Empirical Analysis
by Zhang Chen, Luca Demetrio, Srishti Gupta, Xiaoyi Feng, Zhaoqiang Xia, Antonio Emanuele Cinà, Maura Pintor, Luca Oneto, Ambra Demontis, Battista Biggio, Fabio Roli
First submitted to arxiv on: 14 Jun 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 This research paper investigates the relationship between neural network parameterization and robustness against adversarial examples. Over-parameterized networks, which have been shown to exhibit superior predictive capabilities, have been claimed to be vulnerable to such attacks. However, previous studies have yielded contradictory results, potentially due to flaws in the attack algorithms used. The authors of this paper empirically study the robustness of over-parameterized networks and evaluate the reliability of the attack methods employed. The findings suggest that these networks are indeed robust against adversarial attacks, whereas their under-parameterized counterparts are not. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper looks at how big neural networks do when they’re attacked with special examples designed to trick them. Some people think that these big networks are easy to trick because they have too many parameters. But others say that’s not true. The problem is that the way we test these networks isn’t perfect, which makes it hard to know for sure what’s going on. In this study, the researchers tested how well these big networks do when faced with tricky examples and made sure the tests were reliable. They found that these big networks are actually pretty good at not getting tricked, while smaller networks have a harder time. |
Keywords
* Artificial intelligence * Neural network
