Deep Learning - Machine Learning and Dynamical Systems

Explore cutting-edge research at the intersection of deep learning and dynamical systems through this comprehensive symposium featuring seven expert presentations. Delve into the frequency principle governing deep learning behavior, discover methods for computing Lyapunov functions using neural networks while circumventing dimensional complexity, and examine local conformal autoencoders for advanced data representation. Investigate philosophical connections between Plato's theory of forms and the continuous limits of artificial neural networks, analyze the fundamental tradeoffs between recurrence and attention mechanisms in self-attentive architectures, and understand how control theory principles can ensure AI safety. Learn about innovative applications of neural networks for solving partial differential equations, gaining insights into how machine learning techniques can address complex mathematical and theoretical challenges in dynamical systems.