Physics Informed Neural Networks (PINNs) - Introduction and Applications

Explore the world of Physics Informed Neural Networks (PINNs) in this comprehensive 35-minute video lecture. Delve into the fundamental concept of PINNs, which involves modifying neural networks by incorporating partial differential equations (PDEs) into the loss function to promote solutions that align with known physical principles. Learn about the advantages and disadvantages of this approach, its applications in inference, and discover recommended resources for further study. Examine extensions of PINNs, including Fractional PINNs and Delta PINNs, and understand potential failure modes. Investigate the relationship between PINNs and Pareto fronts. This educational content, produced at the University of Washington with funding support from the Boeing Company, provides a comprehensive overview of PINNs and their applications in physics-informed machine learning.

Intro
PINNs: Central Concept
Advantages and Disadvantages
PINNs and Inference
Recommended Resources
Extending PINNs: Fractional PINNs
Extending PINNs: Delta PINNs
Failure Modes
PINNs & Pareto Fronts
Outro