D2D-Enabled Data Sharing for Distributed Machine Learning at Wireless Network Edge
This work addresses the straggler dilemma in distributed machine learning at the wireless network edge, offering incremental improvements for edge device collaboration.
The paper tackles the straggler problem in mobile edge learning by proposing a device-to-device data sharing approach to adjust computation loads and optimize radio resource allocation, resulting in significantly reduced training delay and improved accuracy for non-IID data.
Mobile edge learning is an emerging technique that enables distributed edge devices to collaborate in training shared machine learning models by exploiting their local data samples and communication and computation resources. To deal with the straggler dilemma issue faced in this technique, this paper proposes a new device to device enabled data sharing approach, in which different edge devices share their data samples among each other over communication links, in order to properly adjust their computation loads for increasing the training speed. Under this setup, we optimize the radio resource allocation for both data sharing and distributed training, with the objective of minimizing the total training delay under fixed numbers of local and global iterations. Numerical results show that the proposed data sharing design significantly reduces the training delay, and also enhances the training accuracy when the data samples are non independent and identically distributed among edge devices.