product

Hamilton–Jacobi equation framework for neural network training