CPS Seminar: Thursday 23 April 2026, 14:00 - 15:00

Building 53, room 4025

Speaker: Dr Aristide Akem, University of Southampton, UK

Title: Machine Learning in network devices: motivation, mechanisms, and applications

Abstract: In-network machine learning (ML) refers to executing trained ML models directly within programmable network devices, such as switches and SmartNICs, using data-plane programming languages like P4. By moving inference into the network, this approach enables line-rate decision making and reduces latency and data movement, opening new possibilities, e.g., for traffic analysis, network security, and routing optimisation.

In this talk, I will introduce the foundations of in-network ML, discuss its motivation and system-level constraints, and examine the mechanisms that make it feasible in modern programmable hardware. I will then highlight representative use cases, ongoing efforts in the area, and outline open challenges and promising directions for future research.

Bio: Dr Aristide Akem is a Lecturer in Computer Science in the Cyber Physical Systems (CPS) group within ECS. Before joining Southampton, he was a postdoctoral researcher in the Computing Infrastructure Group at the University of Oxford. His research spans machine learning, network programming, and mobile networking, as well as related areas such as network security, IoT, and energy. He earned his PhD from Universidad Carlos III de Madrid and IMDEA Networks Institute, following an MSc in Electrical and Computer Engineering from Carnegie Mellon University Africa and an MEng in Telecommunications Engineering from the University of Yaounde I.