Summary of Federated Learning Over Connected Modes, by Dennis Grinwald et al.
Federated Learning over Connected Modes
by Dennis Grinwald, Philipp Wiesner, Shinichi Nakajima
First submitted to arxiv on: 5 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 This paper tackles two major challenges in federated learning: slow global training due to conflicting gradient signals and the need for personalization for local distributions. The authors propose a new approach called Federated Learning over Connected Modes (Floco), which identifies a linearly connected low-loss region in the parameter space of neural networks, known as the solution simplex. Floco assigns clients local subregions within this simplex based on their gradient signals, allowing for personalized model updates that fit local distributions while maintaining global convergence. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Federated learning is a way to train artificial intelligence models together across many devices or computers without sharing private data. But sometimes this process can be slow because different devices have different types of information. To solve this problem, the authors developed a new method called Floco. Floco helps devices learn from each other while still using their own unique information. This makes it faster and more accurate. |
Keywords
* Artificial intelligence * Federated learning
