Summary of Faster Convergence on Heterogeneous Federated Edge Learning: An Adaptive Clustered Data Sharing Approach, by Gang Hu et al.
Faster Convergence on Heterogeneous Federated Edge Learning: An Adaptive Clustered Data Sharing Approach
by Gang Hu, Yinglei Teng, Nan Wang, Zhu Han
First submitted to arxiv on: 14 Jun 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: None
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 Medium Difficulty summary: Federated Edge Learning (FEEL), a pioneering distributed machine learning paradigm, is designed to harness data from Internet of Things (IoT) devices while upholding data privacy. Current FEEL algorithms struggle with non-independent and non-identically distributed (non-IID) data, leading to increased communication costs and compromised model accuracy. To address this issue, we introduce a clustered data sharing framework that selectively shares partial data from cluster heads to trusted associates through sidelink-aided multicasting. The framework’s collective communication pattern is crucial for FEEL training, where both cluster formation and communication efficiency impact training latency and accuracy simultaneously. We decompose the optimization problem into clients clustering and effective data sharing subproblems, proposing a distribution-based adaptive clustering algorithm (DACA) and a stochastic optimization based joint computed frequency and shared data volume optimization (JFVO) algorithm. Experimental results show that our framework facilitates FEEL on non-IID datasets with faster convergence rates and higher model accuracy in limited communication environments. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Low Difficulty summary: This paper is about a new way to learn from lots of small devices like smart home appliances or cars. These devices can share their data without sharing personal information. The problem is that the data might be different, making it harder for the machines to learn together. To solve this issue, we created a system where devices with similar data join “clubs” and share some of their data with other trusted members in the same club. This helps the machines learn more efficiently and accurately. We also developed special algorithms to help these clubs communicate effectively and decide what data to share. |
Keywords
* Artificial intelligence * Clustering * Machine learning * Optimization
