Summary of Towards Universal Certified Robustness with Multi-norm Training, by Enyi Jiang et al.
Towards Universal Certified Robustness with Multi-Norm Training
by Enyi Jiang, David S. Cheung, Gagandeep Singh
First submitted to arxiv on: 3 Oct 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Cryptography and Security (cs.CR)
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 proposes the first multi-norm certified training framework, CURE, which aims to improve union robustness in machine learning models. Current certified training methods can only train models to be robust against a specific type of perturbation, but this new approach allows for better robustness across multiple types of perturbations. The authors develop a theoretical framework to analyze and mitigate the tradeoff between different types of perturbations, and demonstrate that CURE outperforms state-of-the-art certified training methods on various benchmarks, including MNIST, CIFAR-10, and TinyImagenet. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps us make sure our machine learning models are good at working with all kinds of data, not just one specific type. Right now, there are ways to train models that can handle certain types of changes in the data, but these methods don’t work well when the data is changed in different ways. The researchers came up with a new way to train models that can handle many different types of changes at once. They tested their method on some standard datasets and found that it did better than other methods at handling tricky changes. |
Keywords
* Artificial intelligence * Machine learning
