teaching
Courses and mentoring activities.
Physics-Based Machine Learning
Session 1 — Weighted Residual Methods & Foundations for PINNs
Notes covering the mathematical framework behind Physics-Informed Neural Networks: boundary value problems, weighted residual methods (subregion, collocation, Galerkin), neural networks as function approximators, and the PINN loss function. Also includes an introduction to neural operators (DeepONet, FNO, PINO).