Quantum states from normalizing flows

First-principles calculations of non-equilibrium properties of many-body quantum systems via the Hamiltonian formulation in general demand a classical computational resource that scales exponentially in the number of degrees of freedom. One way to avoid this issue is to consider a restricted subspace of the Hilbert space of the system and study time evolution within that subspace, while providing a reliable quantification of uncertainties arising from the reduction of the Hilbert space. Recently, methods of neural network quantum states have been developed as a framework for representing a large subspace of the Hilbert space efficiently with the help of tools from machine learning. In this talk, I will discuss a method for simulating the time evolution of quantum mechanical systems via neural network quantum states with the help of normalizing flows.