Summary of Adversarial Attacks on Hyperbolic Networks, by Max Van Spengler et al.
Adversarial Attacks on Hyperbolic Networks
by Max van Spengler, Jan Zahálka, Pascal Mettes
First submitted to arxiv on: 2 Dec 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
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 paper tackles the issue of adversarial robustness in hyperbolic deep learning, a growing area of research. To address this challenge, the authors propose novel hyperbolic alternatives to traditional FGM and PGD attacks. The proposed methods are evaluated using interpretable synthetic benchmarks and existing datasets, highlighting their differences. Furthermore, the study investigates the variations in adversarial robustness between Euclidean and fully hyperbolic networks, revealing distinct vulnerabilities and limitations of the newly proposed attacks. The findings suggest that models learn unique patterns based on their respective geometries, leading to shifts in adversarial robustness. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about making sure deep learning models are safe from fake or misleading data. Hyperbolic deep learning is a new way of doing things, and the authors want to make sure it’s good at defending against bad data too. They came up with some new ways to test how well models can handle tricky data and found that different types of models have different weaknesses. This means we need to find new ways to make our models better at handling fake or misleading data. |
Keywords
* Artificial intelligence * Deep learning
