Principles of Learning in Distributed Neural Networks

Explore the fundamental principles governing learning in distributed neural networks through this 35-minute lecture by Andrew Saxe from University College London. Delve into the concept of depth as a crucial factor shaping learning dynamics in the brain and mind. Examine mathematical analyses of nonlinear dynamics in simple solvable deep network models. Discover how deep network dynamics can explain individual variations and systematic strategy transitions in rodents learning visual detection tasks over extended periods. Gain insights into the interaction between environmental statistics and nonlinear deep learning dynamics, and their impact on evolving neural representations and behavior during the learning process.