Wireless Edge

Wireless edge computing focuses on bringing computation and data processing closer to end devices, improving latency and bandwidth efficiency for applications like AI model training and inference, extended reality, and media streaming. Current research emphasizes optimizing resource allocation (e.g., through reinforcement learning and Lyapunov optimization), developing efficient model caching strategies (leveraging parameter sharing), and addressing privacy concerns in federated learning through techniques like differential privacy and model sparsification. This field is crucial for enabling the next generation of resource-constrained, latency-sensitive applications across diverse sectors, including IoT, vehicular networks, and satellite communication.

Wireless Edge

Papers

Over-the-Air Ensemble Inference with Model Privacy

Asynchronous Decentralized Learning over Unreliable Wireless Networks

Deep Learning on Multimodal Sensor Data at the Wireless Edge for Vehicular Network