Dynamical Large Deviations of Two-Dimensional Kinetically Constrained Models Using a Neural-Network State Ansatz

We use a neural network ansatz originally designed for the variational optimization of quantum systems to study dynamical large deviations in classical ones. obtain scaled cumulant-generating function activity Fredrickson-Andersen model, prototypical kinetically constrained one and two dimensions, present first size-scaling analysis dimensions. These results provide new route large-deviation functions, highlight broad applicability neural-network state across domains physics.