AI/ML and Physics - Choosing What to Model in Physics-Informed Machine Learning

Explore the first stage of the machine learning process in this 43-minute video lecture on formulating problems to model in AI/ML and Physics. Learn how to incorporate physics into the modeling process and discover new physics through machine learning applications. Dive into various case studies, including super resolution, materials discovery, computational chemistry, digital twins, and shape optimization. Understand the concept of digital twins and discrepancy models, and explore mathematical modeling and chaos in the context of machine learning. Examine benchmark systems and turbulence closure modeling, and gain insights on when not to use machine learning. This comprehensive lecture provides valuable knowledge for researchers and practitioners in the field of Physics Informed Machine Learning.

Intro
Deciding on the Problem
Why do you need an ML Model?
Case Study: Super Resolution
Case Study: Discovering New Physics
Case Study: Materials Discovery
Case Study: Computational Chemistry
Case Study: Digital Twins & Discrepancy Models
Case Study: Shape Optimization
The Digital Twin
Modeling the Math
Modeling the Chaos
Case Study: Climate Modeling
Benchmark Systems
Case Study: Turbulence Closure Modeling
When not to use Machine Learning
Outro