Research Interests

My research considers bandit learning and reinforcement learning - processes in which algorithms learn to take optimal decisions by repeating a process of making a decision, observing data, and updating their model to make better decisions in the future. I am especially interested in settings where the data in question or the model updating processes are complex. My work considers the design and analysis of these algorithms, ensuring methods exist to converge to optimal decisions quickly and reliably. A key feature of my research is collaboration with and inspiration from industry. I am always open to discussing the prospect of new collaborations in theoretical or applied areas.

Web Links

Some further details are available on my personal webpage: https://www.lancaster.ac.uk/staff/grantj/