Title: Structural complexity of snapshots of 2D Fermi-Hubbard systems

Abstract: The development of quantum gas microscopy for two-dimensional optical lattices has provided an unparalleled tool to study the Fermi-Hubbard model (FHM) with ultracold atoms. Spin-resolved projective measurements, or snapshots, have played a significant role in quantifying correlation functions which uncover underlying physical phenomena such as antiferromagnetism at commensurate filling on bipartite lattices, and other charge and spin correlations, as well as dynamical properties at various densities. However, in order to interpret these results and to establish temperature scales, comparison against theory is generally needed. Here, we employ a recent concept, the {\em multi-scale structural complexity}, and show that when computed for the snapshots (of either single spin species, local moments, or total density) it can provide a theory-free property, immediately accessible to experiments. Specifically, after benchmarking results for Ising and XY models, we study the structural complexity of snapshots of the repulsive FHM in the two-dimensional square lattice as a function of doping and temperature. We generate projective measurements using determinant quantum Monte Carlo and compare their complexities against those from the experiment. We demonstrate that these complexities are linked to relevant physical observables such as the entropy and double occupancy. Their behaviors capture the development of correlations and relevant length scales in the system. We provide an open-source code in Python which can be implemented into data analysis routines in experimental settings for the square lattice.