Presenter: Michael Klamkin
Topic: Physics-Informed Neural Networks (PINNs): formulation & pitfalls
This lecture reviews reviews physics-informed neural networks (PINNs), a class of deep learning models that incorporate physical laws into their architecture/training. We will discuss the formulation of PINNs, their applications, and common pitfalls to avoid when using them. Since the field is still developing, we structure the chapter as a review of a few select papers with interesting approaches to applying PINNs for problems in control.
The chapter can be accessed online