A Network-Edge Symbiosis for Accelerated and Distributed AI - PhD

Loughborough University

The Internet of Things (IoT) and next-generation applications like autonomous systems and augmented reality are generating unprecedented volumes of data at the network edge. While Edge AI offers a low-latency alternative to the cloud, it still faces significant bottlenecks. The sheer volume of data can overwhelm network links, and the computational load can saturate edge servers.

This project explores a revolutionary solution that treats the network and edge servers as a single, cohesive computational system. Instead of viewing the network as a simple data pipe, we will transform it into an active partner for the edge, creating a powerful network-edge

symbiosis. The core idea is to intelligently split AI/ML models, executing lightweight, dataintensive tasks like filtering and feature extraction directly in the network using programmable hardware, while reserving more complex inference tasks for the powerful computational

resources at the edge.

This collaborative approach promises to dramatically reduce latency, conserve bandwidth, and create a more efficient, scalable, and resilient infrastructure for distributed AI. This PhD project will be at the forefront of this emerging field, designing and building the foundational

architecture and algorithms for this new generation of intelligent systems.