To be generally useful in the real world, autonomous systems must be able to adapt efficiently to novel environments and tasks - something that approaches such as reinforcement learning struggle with. There is evidence from cognitive science that the impressive adaptivity and efficiency of human learning may be explained in part by our ability to learn socially, by observing and interacting with other people in targeted ways. My project aims to equip autonomous agents with sophisticated social learning abilities by building on ideas from cognitive science and modern computational approaches to learning and decision-making.
I did my undergraduate study at Imperial College London, where I obtained an MEng in molecular bioengineering. Having spent the early years of my course in wet labs, I moved into computational neuroscience for my masters project, which involved designing and implementing a novel information-theoretic algorithm for mapping receptive fields in the mouse visual system. Following a stint as a research assistant in UCL neuroscience, I came to the University of Edinburgh to work at the intersection of autonomous systems and cognitive science. I have also worked as a software engineer at startups in the MedTech and Conversational AI industries.