Summary of Federated Learning with Bilateral Curation For Partially Class-disjoint Data, by Ziqing Fan et al.
Federated Learning with Bilateral Curation for Partially Class-Disjoint Data
by Ziqing Fan, Ruipeng Zhang, Jiangchao Yao, Bo Han, Ya Zhang, Yanfeng Wang
First submitted to arxiv on: 29 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 The proposed FedGELA approach tackles the challenges of partially class-disjoint data in federated learning. By leveraging simplex Equiangular Tight Frame (ETF) for global fairness and local adaptation, FedGELA addresses the angle collapse problem for missing classes and space waste problem for existing classes. The method provides fair classification, avoids inaccurate updates, and utilizes locally missing class spaces. Experimental results on various datasets demonstrate an averaged improvement of 3.9% over FedAvg and 1.5% over best baselines, with local and global convergence guarantees. The authors provide open-source code at https://github.com/MediaBrain-SJTU/FedGELA.git. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Partially class-disjoint data in federated learning is a challenge that existing methods can’t overcome. A new approach called FedGELA addresses this issue by using a simplex Equiangular Tight Frame to make the classifier fair and equal for all classes. This helps avoid inaccurate updates and makes the most of missing class spaces. The results show that FedGELA works well, improving performance by 3.9% over another method and 1.5% over other baselines. |
Keywords
» Artificial intelligence » Classification » Federated learning
