Summary of Hierarchical Federated Learning in Multi-hop Cluster-based Vanets, by M. Saeid Haghighifard and Sinem Coleri
Hierarchical Federated Learning in Multi-hop Cluster-Based VANETs
by M. Saeid HaghighiFard, Sinem Coleri
First submitted to arxiv on: 18 Jan 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Distributed, Parallel, and Cluster Computing (cs.DC); Networking and Internet Architecture (cs.NI); Systems and Control (eess.SY)
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 proposes a novel framework for hierarchical federated learning (HFL) in Vehicular Ad hoc Networks (VANET). Federated learning allows for training machine learning models on local data without transmitting the data itself, reducing transmission overhead and protecting user privacy. However, implementing FL in VANETs poses challenges due to limited communication resources, high vehicle mobility, and statistical diversity of data distributions. The proposed HFL framework uses a weighted combination of average relative speed and cosine similarity as a clustering metric to consider both data diversity and high vehicle mobility. This ensures convergence with minimum changes in cluster heads while tackling non-independent and identically distributed (non-IID) data scenarios. The framework also includes mechanisms for seamless transitions of cluster heads, transferring the most recent FL model parameter, and merging cluster heads to reduce their count and associated overhead. Simulation results demonstrate that HFL improves accuracy and convergence time significantly while maintaining an acceptable level of packet overhead. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Federated learning in Vehicular Ad hoc Networks (VANET) helps reduce transmission overhead and protect user privacy. But implementing it is tricky because of limited communication resources, moving vehicles, and different types of data. To solve this problem, researchers proposed a new way to group vehicles together using a special metric that considers both the speed of the vehicles and how similar their data is. This approach allows the model to converge quickly without changing the leader of each group too often. The framework also includes ways to switch between leaders smoothly and combine groups to reduce overhead. The results show that this new approach works well, improving accuracy and speed while keeping packet overhead low. |
Keywords
* Artificial intelligence * Clustering * Cosine similarity * Federated learning * Machine learning
