Skip to main content

Computer Science

Dr George De Ath

Research Fellow
Computer Science

I am a Proleptic Lecturer in Machine Learning and Optimisation, contributing as a co-Investigator on Project Bluebird, a prosperity partnership programme between the University of Exeter, NATS and The Alan Turing Institute. Within this programme, I co-lead the agent development research theme, which focuses on designing advanced AI agents to manage air traffic control tasks both safely and efficiently. Following a well-recieved presentation at the British Science Festival, our work was recently featured in the Financial Times.

 

Prior to my current role, I was a permanent Research Fellow for the Institute for Data Science and Artificial Intelligence (IDSAI) at the University of Exeter. Between 2019 and 2021, I worked as a Postdoctoral Research Fellow on two UKRI-funded projects, RIBA to Reality: Deep Digital Twin to enable Human-Centric Buildings for a Carbon Neutral Future and Rapid Calibration for Operational and Strategic Digital Twins. I hold both an MSci in Computer Science and Mathematics and a Ph.D. in Computer Science from the University of Exeter.

 

Research Interests

My research focuses on real-world applications of machine learning and optimisation. A key area of my work involves developing novel approaches for air traffic control and broader air traffic management challenges. My primary interests include the optimisation (calibration) of expensive-to-evaluate problems (models) using Bayesian optimisation, as well as more general single- and multi-objective optimisation tasks, and solving regression/classification problems in machine learning. I have experience in evolutionary optimisation, probabilistic modelling, with a particular focus on Gaussian processes, uncertainty quantification, and general machine learning approaches, e.g., neural networks, random forests, gradient boosted machines, etc.

 

Prospective Students

I welcome MSc and PhD students interested in (Bayesian) optimisation, machine learning, and uncertainty quantification, including both traditional statistical techniques and deep learning approaches. I also encourage students with overlapping interests to reach out. Current opportunities include:

 

Current Students

  • (PhD) Ben Carvell: Safety-Critical Decision Making under Uncertainty – Machine Learning Approaches for Tactical Air Traffic Control
  • (PhD) James Edwards: Transforming Air Traffic Control: Sequence Prediction and Modelling with Large Language Models
  • (MRes) Lawrence Knight: Extending Pluribus to use Deep Counterfactual Regret Minimisation

View full profile