Summary of Generalization Bound and New Algorithm For Clean-label Backdoor Attack, by Lijia Yu and Shuang Liu and Yibo Miao and Xiao-shan Gao and Lijun Zhang
Generalization Bound and New Algorithm for Clean-Label Backdoor Attack
by Lijia Yu, Shuang Liu, Yibo Miao, Xiao-Shan Gao, Lijun Zhang
First submitted to arxiv on: 2 Jun 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Cryptography and Security (cs.CR); Statistics Theory (math.ST)
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 The paper addresses the generalizability of learning methods in the context of backdoor attacks, which are a type of data poisoning. The authors establish algorithm-independent generalization bounds for clean-label backdoor attacks, providing upper bounds for population errors in terms of empirical error on the poisoned training dataset. The paper also proposes a new clean-label backdoor attack that combines adversarial noise and indiscriminate poison, demonstrating its effectiveness in various settings. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper explores how to measure the reliability of learning methods when they’re used on data that’s been tampered with. This is important because it can help us figure out if our models will work well even when the data is messed up. The authors come up with some new math formulas that show how well a model will do in different situations, and they also propose a new way to make backdoor attacks that combines two types of fake signals. |
Keywords
» Artificial intelligence » Generalization
