Highly efficient cloud computing via Adaptive Hierarchical Federated Learning
Shao Dan, Qinyi Zhu · Optimizations in Applied Machine Learning · 2025
Cloud computing has revolutionized the way data is processed and stored, leading to increased demand for efficient machine learning models. However, the current centralized nature of cloud-based machine learning poses challenges in terms of scalability and privacy protection. This paper addresses these obstacles by proposing a novel approach called Adaptive Hierarchical Federated Learning. This approach enables the efficient distribution of machine learning tasks across multiple layers of a hierarchical cloud architecture, allowing for improved scalability and enhanced privacy preservation. The innovative method presented in this paper harnesses the power of federated learning while adapting dynamically to the varying computational resources within the hierarchical cloud environment. Through extensive experiments, the effectiveness and efficiency of the proposed Adaptive Hierarchical Federated Learning are demonstrated, highlighting its potential to significantly advance the field of cloud computing.