Summary of Differentially Private Worst-group Risk Minimization, by Xinyu Zhou et al.
Differentially Private Worst-group Risk Minimization
by Xinyu Zhou, Raef Bassily
First submitted to arxiv on: 29 Feb 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); 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 We initiate a systematic study on worst-group risk minimization under (, )-differential privacy (DP). Our goal is to privately find a model that approximately minimizes the maximal risk across p sub-populations with different distributions. We present a new algorithm achieving excess worst-group population risk of ( + ), nearly optimal when each distribution is observed via a fixed-size dataset of size K/p. Our result is based on a new stability-based analysis for the generalization error. We also propose an algorithmic framework using any DP online convex optimization algorithm as a subroutine, with an excess risk bound of ( + ). Finally, we study differentially private worst-group empirical risk minimization in the offline setting, with a new algorithm achieving nearly optimal excess risk of (). |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about making sure that models work well for all groups of people, even if some groups are very different. We want to do this while keeping the data private and protected. The authors present a new way to achieve this using an algorithm that can find good models quickly. They also show how their method works in practice by testing it on real-world data. |
Keywords
* Artificial intelligence * Generalization * Optimization
