Novel Neural Network Architectures for Simulations of Quantum Field Theories

Machine learning methods have been suggested as alternatives to the standard algorithms used for simulations of lattice field theories. In this talk, I will present new neural network architectures inspired by effective field theories, designed to improve the scaling of the training cost for the generation of lattice field theory configurations. We first address poor acceptance rates in simulations of large lattices for scalar field theory in two dimensions and then discuss possible extensions to gauge theories in higher dimensions.