Summary of Maverick-aware Shapley Valuation For Client Selection in Federated Learning, by Mengwei Yang et al.
Maverick-Aware Shapley Valuation for Client Selection in Federated Learning
by Mengwei Yang, Ismat Jarin, Baturalp Buyukates, Salman Avestimehr, Athina Markopoulou
First submitted to arxiv on: 21 May 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: 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 address a key challenge in Federated Learning (FL) systems, specifically the issue of data heterogeneity when clients with rare data are present. These “Maverick” clients have exclusive ownership of one or more data classes, and their participation is crucial for improving model performance. The authors design a Maverick-aware Shapley valuation to fairly evaluate the contribution of these clients, class-wise per label. They also propose FedMS, a client selection mechanism that intelligently selects clients based on their contribution score, computed using the Maverick-aware Shapley values. Compared to various baselines, FedMS achieves better model performance and fairer Shapley rewards distribution. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Federated Learning is a way for computers to work together without sharing their private data. One problem with this system is that some devices have very unique data, which makes it hard for the model to work well. These “Maverick” devices are important for making the model better. The researchers in this paper developed a new way to calculate how much each device contributes to the model, and they also created an algorithm to choose which devices to use next. This helps make sure that all devices are treated fairly and that the model gets better. |
Keywords
» Artificial intelligence » Federated learning
