Organizing Knowledge for Flexible Behavior in Neural Networks

Watch a Stanford Wu Tsai Neurosciences Institute seminar exploring how animals develop flexible behavior through neural representations and knowledge generalization. Delve into groundbreaking research presented by Sir Henry Wellcome Scholar James Whittington, who splits his time between Stanford University and Oxford University. Examine a hippocampal model for learning and generalizing building blocks that accounts for various cell types in the hippocampus and entorhinal cortex, drawing connections to transformer neural networks. Explore theoretical findings about neural population representations and their relationship to grid cell warping. Investigate the normative theory of entorhinal grid cells and how individual building blocks should be represented in the brain. Part of the Stanford Center for Mind, Brain, Computation and Technology (MBCT) seminar series, this 75-minute presentation advances understanding of computational approaches in neuroscience research and their implications for flexible behavior.