Computation and Communication Efficient Lightweighting Vertical Federated Learning

--The exploration of computational and communication efficiency within Federated Learning (FL) has emerged as a prominent and crucial field of study. While most existing efforts to enhance these efficiencies have focused on Horizontal FL, the distinct processes and model structures of V ertical FL preclude the direct application of Horizontal FLbased techniques. In response, we introduce the concept of Lightweight V ertical Federated Learning (L VFL), targeting both computational and communication efficiencies. This approach involves separate lightweighting strategies for the feature model, to improve computational efficiency, and for feature embedding, to enhance communication efficiency. Moreover, we establish a convergence bound for our L VFL algorithm, which accounts for both communication and computational lightweighting ratios. Our evaluation of the algorithm on a image classification dataset reveals that L VFL significantly alleviates computational and communication demands while preserving robust learning performance. This work effectively addresses the gaps in communication and computational efficiency within V ertical FL. Federated learning (FL) is a distributed machine learning paradigm that enables a set of clients with decentralized data to collaborate and learn a shared model under the coordination of a centralized server. In FL, data is stored on edge devices in a distributed manner, which reduces the amount of data that needs to be uploaded and decreases the risk of user privacy leakage. The majority of prior research focuses on horizontal federated learning (HFL), where participants share the same feature space but have distinct sample spaces.