Summary of Attacking Bayes: on the Adversarial Robustness Of Bayesian Neural Networks, by Yunzhen Feng et al.
Attacking Bayes: On the Adversarial Robustness of Bayesian Neural Networks
by Yunzhen Feng, Tim G. J. Rudner, Nikolaos Tsilivis, Julia Kempe
First submitted to arxiv on: 27 Apr 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Methodology (stat.ME); 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 research paper examines the claim that Bayesian neural networks (BNNs) are inherently robust to adversarial perturbations. The study investigates whether state-of-the-art BNN inference methods can be successfully broken using relatively unsophisticated attacks for three tasks: label prediction, adversarial example detection, and semantic shift detection. The results show that BNNs trained with approximate inference methods or Hamiltonian Monte Carlo are highly susceptible to adversarial attacks. Furthermore, the study identifies conceptual and experimental errors in previous works claiming inherent robustness of BNNs. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Bayesian neural networks (BNNs) are a type of artificial intelligence that uses probability theory to make predictions. Some researchers claim that these networks are naturally good at dealing with tricky data that’s been altered or tampered with. But is this really true? In this study, scientists looked into whether BNNs can withstand attempts to trick them using simple attacks. They found out that even the best BNNs aren’t as robust as we thought they were. |
Keywords
» Artificial intelligence » Inference » Probability
