Stable Adaptation and Learning in Nonlinear Systems and Neural Networks

Explore the fascinating intersection of neuroscience, robotics, and machine learning in this 35-minute lecture by Jean-Jacques Slotine from MIT. Delve into the remarkable efficiency of the human brain compared to artificial systems, and discover how modern nonlinear systems tools can provide valuable insights into collective computation and learning in large dynamical networks. Examine the potential of stable implicit sparse regularization in adaptive prediction and control for selecting relevant dynamic models. Learn how Riemannian contraction can offer more generalized results than traditional gradient descent methods based on convexity. Gain a deeper understanding of the challenges and opportunities in bridging the gap between biological and artificial intelligence through this thought-provoking presentation at the Institut des Hautes Etudes Scientifiques (IHES).