Summary of Variational Randomized Smoothing For Sample-wise Adversarial Robustness, by Ryo Hase et al.
Variational Randomized Smoothing for Sample-Wise Adversarial Robustness
by Ryo Hase, Ye Wang, Toshiaki Koike-Akino, Jing Liu, Kieran Parsons
First submitted to arxiv on: 16 Jul 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Machine Learning (stat.ML)
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 abstract presents a novel approach to defending neural networks against adversarial attacks, specifically targeting small input perturbations that degrade model performance. The proposed randomized smoothing technique introduces a noise level selector, allowing for per-sample noise levels tailored to each input. This framework enhances empirical robustness against adversarial attacks and provides certified robustness for sample-wise smoothing methods. Experimental results demonstrate the effectiveness of this approach. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper shows how to make neural networks more secure by using random noise in a special way. Adversarial examples are small changes to inputs that can cause models to fail. To fix this, the researchers propose a new method called randomized smoothing with a “noise level selector”. This lets each input get its own amount of random noise, which helps defend against attacks. The results show that this method makes neural networks more robust and provides extra guarantees about their performance. |
