I research reinforcement learning with graph based methods in the UKRI CDT for Accountable, Responsible, and Transparent AI, under the supervision of Prof Özgür Şimşek.
The complex graphical structure generated by interactions between a reinforcement learning agent and its environment is a rich source of additional information that can be exploited for a number of aims. My research combines graph theory and reinforcement learning to achieve more efficient value function approximation, and for explaining a policy the agent is pursuing.
Reinforcement Learning.
Explainable AI.
BSc Mathematics with Finance, University of Brighton.
MSc Mathematics, University of Leeds.
MSc Computer Science, University of Edinburgh.
Industry experience in Operations Research.
Prof Özgür Şimşek
Dr Janina Hoffmann
