Summary of Tackling Selfish Clients in Federated Learning, by Andrea Augello et al.
Tackling Selfish Clients in Federated Learning
by Andrea Augello, Ashish Gupta, Giuseppe Lo Re, Sajal K. Das
First submitted to arxiv on: 22 Jul 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 A novel attack on Federated Learning (FL) is proposed in this paper, where some intelligent clients deliberately deviate from the standard training process to prioritize their local data distribution. This misbehavior is referred to as selfishness, and it can significantly decrease the accuracy of the global model even when only a small percentage of clients behave selfishly. To mitigate this effect, the authors propose a Robust aggregation strategy for FL server, called RFL-Self, which incorporates an innovative method to recover the true updates of selfish clients using robust statistics. The experimental results on MNIST and CIFAR-10 datasets demonstrate that RFL-Self can effectively mitigate the impact of selfishness without degrading the global model performance. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper discusses a problem with Federated Learning, where some devices try to make the overall model better suited to their own data rather than working together. This makes it less accurate and can even decrease its accuracy by up to 36% if only 2% of devices do this. The authors suggest a way to fix this called RFL-Self, which helps the main server understand what the selfish devices are trying to do and adjust accordingly. |
Keywords
» Artificial intelligence » Federated learning
