The papers in this special section focus on distributed signal processing for edge learning (EL). EL is a new and promising technology for implementing artificial intelligence (AI) algorithms at edge devices over wireless networks. With the explosive growth in global data traffic, the number of edge devices such as mobile edge computing (MEC), satellite networks, and the Internet-ofthings (IoT) devices increase rapidly. Machine learning (ML) techniques, including deep learning (DL), federated learning (FL), and reinforcement learning (RL), are effective approaches to improve the performance and efficiency of edge networks.
Edge learning (EL) is a promising technology for implementing artificial intelligence (AI) algorithms at edge devices over wireless networks. With the explosive growth in global data traffic, the number of edge devices such as mobile edge computing (MEC), satellite networks, and the Internet-of-things (IoT) devices increase rapidly. Machine learning (ML) techniques, including deep learning (DL), federated learning (FL), and reinforcement learning (RL), are effective approaches to improve the performance and efficiency of edge networks. Traditionally, the AI model in communication systems is trained on a centralized cloud server, which may suffer from the problem of large computational complexity and heavy signaling overhead. Since EL deploys computation-intensive model training at the network edge, highly-distributed real-time data generated by edge devices can be accessed rapidly to accelerate AI training. However, edge devices with limited computational and storage resources lead to new challenges in resource allocation. Other open challenges include data privacy protection and communication latency control. A detailed introduction to distributed EL is given in paper [A1], entitled “Edge Learning for B5G Networks with Distributed Signal Processing: Semantic Communication, Edge Computing, and Wireless Sensing,” which provides a comprehensive overview of practical distributed EL techniques and their interplay with advanced wireless communication optimization designs.
