Four SoM researchers appointed Fellows of Edinburgh’s Generative AI Lab

Martin Brolly, a Postdoctoral Research Associate, is developing hybrid models that couple generative machine learning models to dynamical models. He aims to accelerate simulations and accurately quantify uncertainty in model predictions, with applications in Earth system modelling.

Miguel de Carvalho, Chair of Statistical Data Science, is developing models for the complex sequential nature of extreme events, as well as generative AI models for extreme events. In collaboration with industry, his work bridges statistical models with large language models to interface quantitative and language-based aspects. Miguel is funded by Abrdn, GAIL, Prob_AI, and the Royal Society of Edinburgh.

Professor of Numerical Analysis Desmond Higham’s research in the area of generative AI is around understanding and, where possible, mitigating the inherent vulnerabilities in AI systems, especially with respect to attacks by adversaries. He also produces theory and tools that inform AI developers and regulators who are interested in safety and security.

Reader in Machine Learning Francesco Tudisco’s research focuses on generative AI for scientific simulations. He has worked on designing foundation models for the Sun’s atmosphere and is actively working on generative models to forecast partial differential equations and other complex systems. He is also involved in researching how to enhance the robustness and trustworthiness of large AI models, as well as improve the efficiency of training, fine-tuning, and inference.