Physics Informed Machine Learning: High-Level Overview of AI and ML in Science and Engineering

Explore the integration of physics into machine learning through this comprehensive 47-minute video lecture. Discover how to incorporate physical knowledge into each of the five stages of the machine learning process: problem formulation, data collection and curation, architecture selection, loss function design, and optimization algorithm implementation. Learn about the importance of physics-informed machine learning in engineering applications, especially for systems governed by physical laws and safety-critical components. Understand how this approach enables more effective learning from sparse and noisy datasets. Dive into case studies, including encoding pendulum movement, and gain insights into physics-informed problem modeling, data curation, architecture design, loss functions, and optimization algorithms. Prepare for an in-depth exploration of AI and machine learning applications in science and engineering.

Intro
What is Physics Informed Machine Learning?
Case Study: Encoding Pendulum Movement
The Five Stages of Machine Learning
A Principled Approach to Machine Learning
Physics Informed Problem Modeling
Physics Informed Data Curation
Physics Informed Architecture Design
Physics Informed Loss Functions
Physics Informed Optimization Algorithms
What This Course Will Cover
Outro