Summary of Gpfl: a Gradient Projection-based Client Selection Framework For Efficient Federated Learning, by Shijie Na et al.
GPFL: A Gradient Projection-Based Client Selection Framework for Efficient Federated Learning
by Shijie Na, Yuzhi Liang, Siu-Ming Yiu
First submitted to arxiv on: 26 Mar 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 The proposed GPFL framework addresses limitations in existing methods for federated learning client selection by measuring client value based on local and global descent directions. This approach enhances model accuracy and communication efficiency while handling data heterogeneity and computational burdens. Experimental results demonstrate a 9% improvement in test accuracy on the FEMINST dataset, with shorter computation times achieved through pre-selection and parameter reuse. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Federated learning helps devices learn together without sharing their data. Choosing which devices to use is important for making sure the model works well and doesn’t take too much time or effort. The GPFL framework does this by comparing how well each device’s data fits with the overall model. This helps make the model more accurate and efficient. Tests on two different datasets show that GPFL is better than other methods, especially when devices have very different types of data. |
Keywords
* Artificial intelligence * Federated learning
