Tutorial: A Statistical Tour of Physics-Informed Machine Learning (PIML)

This tutorial aims to provide a statistical perspective on physics-informed learning methods, with a focus on both theoretical understanding and practical implementation.

Lecturers

Claire Boyer
Professor, Université Paris-Saclay

Nathan Doumèche
PhD Candidate, EDF & Sorbonne Université

Materials

Lecture notes (handwritten, pdf) (and additional slides)
Jupyter notebooks for hands-on sessions (via Google Colab)
Python package: pikernel

Schedule

Day	Time	Program
Tuesday		Lectures I & II: A Statistical Perspective on PINNs Lecture notes Additional slides
Wednesday		Practical session I: Implementing PINNs Notebook: [.ipynb] [display] Correction: [.ipynb] [display] Lecture III: A primer on kernel methods (Part I) Lecture notes
Thursday		Lecture IV: A primer on kernel methods (Part II) Lecture notes
Friday		Practical session II: Implementing kernel methods Notebook: [.ipynb] [display] Correction: [.ipynb] [display] Lecture V: Physics-informed kernel learning Lecture notes Additional slides
Saturday		Practical session III: Implementing physics-informed kernels methods Notebook: [.ipynb] [display] Correction: [.ipynb] [display]

Using Google Colab

1. Log in with your Google Account

2. Introduction to Google Colab

This introductory notebook provides an overview of Colab's features.