Summary of From Optimization to Generalization: Fair Federated Learning Against Quality Shift Via Inter-client Sharpness Matching, by Nannan Wu et al.
From Optimization to Generalization: Fair Federated Learning against Quality Shift via Inter-Client Sharpness Matching
by Nannan Wu, Zhuo Kuang, Zengqiang Yan, Li Yu
First submitted to arxiv on: 27 Apr 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computer Vision and Pattern Recognition (cs.CV)
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 Federated learning with Inter-client Sharpness Matching (FedISM) method addresses the issue of fairness in federated learning when dealing with decentralized medical data. In a scenario where imaging quality varies across institutions due to equipment malfunctions, FedISM enhances local training and global aggregation by incorporating sharpness-awareness. This allows for harmonization of sharpness levels across clients, ensuring fair generalization. The approach is evaluated using the ICH and ISIC 2019 datasets, demonstrating superiority over current state-of-the-art federated learning methods in promoting fairness. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Federated learning is a way to train AI models with medical data from different hospitals or clinics without sharing the data itself. This helps protect patient privacy. But sometimes, the quality of images used for training can be very different at each hospital. This can cause the model to become unfair, favoring hospitals with better equipment. Researchers have developed a new method called FedISM that tries to fix this problem by making sure all hospitals contribute equally to the model’s learning. |
Keywords
» Artificial intelligence » Federated learning » Generalization
