Lecture: Automatic Differentiation & Adjoint Methods in Differentiable Physics

As part of our master course in Advanced Deep Learning for Physics (IN2298), I gave a lecture on autodiff and adjoint methods. You can find the lecture slides here. The lecture was recorded and is available on YouTube:

In it, I cover:

1. A functional (JAX/Julia-inspired) viewpoint on autodiff in terms of Jvp/Pushforward and vJp/Pullback
2. Identifying hierarchy levels in autodiff (scalar-mode, vector-mode, continuous-mode)
3. A comparison of Optimize-then-Discretize (OtD) vs. Discretize-then-Optimize (DtO)
4. Special aspects of differentiable physics like differentiating over linear and nonlinear solvers
5. Advanced topics and recent research directions