Physics-Enhanced Gaussian Processes for Learning of Electromechanical Systems

Explore a comprehensive lecture on physics-enhanced Gaussian processes for learning electromechanical systems. Delve into the integration of physical principles with data-driven approaches to model nonlinear systems more efficiently and accurately. Discover the innovative Gaussian Process Port-Hamiltonian systems (GP-PHS) as a physics-informed, nonparametric Bayesian learning approach with uncertainty quantification. Learn about recent advancements in physics-enhanced variational autoencoders that utilize physically enhanced Gaussian process priors for improved efficiency and physically correct predictions. Gain insights into the benefits of incorporating physical prior knowledge in machine learning models, demonstrated through simulations of oscillating particles. Presented by Thomas Beckers, Assistant Professor of Computer Science at Vanderbilt University, this talk offers valuable knowledge for researchers and practitioners in the fields of machine learning, control systems, and electromechanical engineering.

DDPS | Physics-enhanced Gaussian Processes for Learning of Electromechanical Systems |Thomas Beckers