Overview
Explore the Fourier Neural Operator (FNO) in this 18-minute video lecture on Physics Informed Machine Learning. Delve into concepts such as operators as images, Fourier as convolution, and zero-shot super resolution. Examine the generalization of neural operators, conditions and operator kernels, and mesh invariance. Understand the advantages of neural operators over other methods and discover their applications in Green's Function and Laplace Neural Operators. Produced at the University of Washington with funding support from the Boeing Company, this comprehensive overview provides valuable insights into advanced machine learning techniques for physics-based problems.
Syllabus
Intro
Operators as Images, Fourier as Convolution
Zero-Shot Super Resolution
Generalizing Neural Operators
Conditions and Operator Kernels
Mesh Invariance
Why Neural Operators // Or Neural operators vs other methods
Result: Green's Function
Laplace Neural Operators
Outro
Taught by
Steve Brunton