Summary of Personalized Federated Learning Via Admm with Moreau Envelope, by Shengkun Zhu et al.
Personalized Federated Learning via ADMM with Moreau Envelope
by Shengkun Zhu, Jinshan Zeng, Sheng Wang, Yuan Sun, Zhiyong Peng
First submitted to arxiv on: 12 Nov 2023
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 federated learning approach, FLAME, achieves sublinear convergence rates and alleviates the need for hyperparameter tuning by leveraging an alternating direction method of multipliers (ADMM). This allows for efficient training on heterogeneous data. FLAME also introduces a biased client selection strategy to accelerate model convergence. Theoretical analysis demonstrates global convergence under both unbiased and biased strategies. Experimental results show that FLAME outperforms state-of-the-art methods in terms of model performance, with an average speedup of 3.75x in communication efficiency. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Federated learning is a way for devices to work together on a shared task without sharing their own data. The problem is that when the devices have different types of data, it can be hard for them to agree on what the best model is. FLAME is a new approach that makes it easier for these devices to come to an agreement by using a special kind of math called ADMM. This helps the devices learn from each other without needing to adjust lots of complicated settings. FLAME also has a way to pick which devices should be in charge, which can make the learning process faster. |
Keywords
* Artificial intelligence * Federated learning * Hyperparameter
