A neural network f(x,y) defines a vector field dx/dt = f₁(x,y), dy/dt = f₂(x,y). Train on target trajectories and watch the phase portrait evolve.