Efficient Federated Learning Via Low-Rank Gradient Compression for Intelligent Transportation System

Within the realm of intelligent transportation systems, be it for autonomous vehicles or other applications, the window of opportunity for executing distributed learning is constrained. Hence, the imperative lies in devising expeditious convergence techniques for pertinent models as a pivotal technical strategy for tackling this challenge. Recent research has homed in on gradient compression, hastening model convergence, and curbing communication expenses. This study presents a swift convergence approach for federated learning, employing low-rank matrix factorization methods. Empirical findings attest to the efficacy of this approach in amplifying the convergence rate of federated learning.