Machine Learning Detection of Correlations in Snapshots of Ultracold Atoms in Optical Lattices

TL;DR

This paper enhances neural network interpretability for classifying ultracold atom snapshots, revealing new fermion arrangements at low temperatures and doping levels in optical lattices.

Contribution

It introduces a method to analyze multiple convolutional filters, improving understanding of learned patterns in quantum gas microscopy data.

Findings

Identified physically relevant patterns in neural network filters.

Discovered new fermion arrangements at low temperatures and doping.

Validated approach at half-filling with known antiferromagnetic correlations.

Abstract

Recent proposals have suggested the use of supervised learning with convolutional neural networks to shed light on some of the less well known phases of the Fermi-Hubbard model through the classification of snapshots from the quantum gas microscopy of ultracold atoms in optical lattices. However, there have been challenges in the interpretability of networks with more than one convolutional filter coupled to the input images. Here, we expand on previous work by considering multiple filters in the first convolutional layer and developing a process for analyzing the physical relevance of patterns obtained in the trained filters. We benchmark our approach at half-filling, where strong antiferromagnetic correlations are known to be present, and we find that upon hole doping, previously unknown patterns arise at temperatures below the tunneling amplitude. These patterns may be a signature of interesting arrangements of fermions in the lattice.