Supervised Training of Neural-Network Quantum States for the Next Nearest Neighbor Ising model

TL;DR

This paper compares supervised training strategies for neural network quantum states applied to the next-nearest neighbor Ising model, highlighting the effectiveness of the overlap loss function across different phases.

Contribution

It systematically evaluates various neural network architectures, hyper-parameters, and loss functions for quantum state representation in the next-nearest neighbor Ising model.

Findings

Overlap loss function improves training across all phases.

Rescaling neural networks enhances performance with overlap loss.

Different phases require tailored training strategies.

Abstract

Different neural network architectures can be unsupervisedly or supervisedly trained to represent quantum states. We explore and compare different strategies for the supervised training of feed forward neural network quantum states. We empirically and comparatively evaluate the performance of feed forward neural network quantum states in different phases of matter for variants of the architecture, for different hyper-parameters, and for two different loss functions, to which we refer as \emph{mean-squared error} and \emph{overlap}, respectively. We consider the next-nearest neighbor Ising model for the diversity of its phases and focus on its paramagnetic, ferromagnetic, and pair-antiferromagnetic phases. We observe that the overlap loss function allows better training of the model across all phases, provided a rescaling of the neural network.