Evaluation of parameterized quantum circuits: on the relation between classification accuracy, expressibility, and entangling capability

Abstract An active area of investigation in the search for quantum advantage is machine learning. Quantum learning, and parameterized circuits a hybrid quantum-classical setup particular, could bring advancements accuracy by utilizing high dimensionality Hilbert space as feature space. But ability circuit to uniformly address good indicator classification accuracy? In our work, we use methods quantifications from prior art perform numerical study order evaluate level correlation. We find moderate strong correlation between achieved that entail single embedding layer followed 1 or 2 designs. This based on encompassing 19 both 1- 2-layer configurations, evaluated 9 datasets increasing difficulty. also entangling capability similar setup, weak Future work will focus evaluating if this holds different