James Whittington

I try to understand the neural mechanisms underlying flexible behaviour, focusing on tasks in which structural knowledge must be learned and generalised. For example, understanding that presents should be wrapped before being opened, uses similar structured knowledge as knowing that you can only vote if you are already registered. Such structured knowledge, and their corresponding tasks, are often hierarchical, compositional, and require inferences. By combining modelling and theory, I investigate the optimal neural representations and algorithms for solving these tasks and relate them to prefrontal cortex and the hippocampal formation. I make use of neural networks trained to learn and generalise on structured tasks, as well as analysing these models theoretically to provide a mathematical formalism for optimal neural representations and algorithms that can explain when and why neural systems learn different representations and algorithms. Lastly, I use the understanding gained from neuroscience tasks to build bridges between natural and artificial intelligence.

I have spent a long time at Oxford with a undersgraduate in physics, then medical degree, and then a DPhil in Neuroscience. I did my postdoc (Sir Henry Wellcome Fellow) at Stanford. I have worked at AI startups as well as big tech. I am also a co-founder of Thinking About Thinking, organising its scientific agenda as well as the programmes for several Summits and conferences each year.