Summary of Sok: Challenges and Opportunities in Federated Unlearning, by Hyejun Jeong et al.
SoK: Challenges and Opportunities in Federated Unlearning
by Hyejun Jeong, Shiqing Ma, Amir Houmansadr
First submitted to arxiv on: 4 Mar 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC)
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 In this paper, researchers explore how federated learning (FL) can be adapted to accommodate the growing need for model owners to “forget” previously learned data, a concept known as machine unlearning. Federated learning allows multiple parties to jointly train models without sharing their individual data, respecting privacy regulations. However, emerging requirements may necessitate the ability to selectively forget learned information. The authors highlight the unique challenges in applying existing unlearning techniques from centralized settings to FL, which requires interactivity, stochasticity, heterogeneity, and limited accessibility. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Machine learning researchers are working on a new way to make AI models forget certain things they’ve learned. This is called “machine unlearning.” It’s important because laws like GDPR and CPRA require companies to respect people’s privacy. One type of machine unlearning is called federated learning (FL). FL lets many companies work together on an AI model without sharing their individual data. But now, there’s a need for models to be able to forget things they’ve learned. This is harder than it seems because the way AI models learn in FL is different from how they learn in other settings. |
Keywords
* Artificial intelligence * Federated learning * Machine learning
