Summary of Federated Learning with Relative Fairness, by Shogo Nakakita et al.
Federated Learning with Relative Fairness
by Shogo Nakakita, Tatsuya Kaneko, Shinya Takamaeda-Yamazaki, Masaaki Imaizumi
First submitted to arxiv on: 2 Nov 2024
Categories
- Main: Machine Learning (stat.ML)
- Secondary: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
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 framework aims to achieve relative fairness for clients, unlike traditional frameworks that prioritize absolute fairness. The new approach uses a minimax problem formulation to minimize relative unfairness, building upon distributionally robust optimization (DRO) methods. A novel fairness index is introduced to assess and improve the relative fairness of trained models. Theoretical guarantees demonstrate consistent reductions in unfairness. An algorithm called Scaff-PD-IA balances communication and computational efficiency while maintaining minimax-optimal convergence rates. Empirical evaluations on real-world datasets confirm its effectiveness in reducing disparity while maintaining model performance. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper proposes a new way to make sure that different groups of people get the same benefits from using artificial intelligence models. Right now, many AI systems are biased because they’re trained on data that’s not fair or representative of everyone’s experiences. The researchers developed a special kind of “fairness” math problem to help solve this issue. They also created an algorithm called Scaff-PD-IA that can be used to train AI models in a way that’s more fair and equal. This is important because it helps ensure that people who need AI the most get the same benefits as everyone else. |
Keywords
» Artificial intelligence » Federated learning » Optimization
