qLEET: Visualizing Loss Landscapes, Expressibility, Entangling Power and Training Trajectories for Parameterized Quantum Circuits

TL;DR

qLEET is an open-source Python tool for analyzing parameterized quantum circuits, enabling visualization of loss landscapes, and assessing expressibility and entangling power to improve quantum algorithms.

Contribution

The paper introduces qLEET, a novel software package that provides comprehensive analysis and visualization tools for PQCs, supporting multiple quantum computing frameworks.

Findings

qLEET effectively visualizes high-dimensional loss landscapes.

It quantifies expressibility and entangling power of PQCs.

The tool aids in designing better hybrid quantum-classical algorithms.

Abstract

We present qLEET, an open-source Python package for studying parameterized quantum circuits (PQCs), which are widely used in various variational quantum algorithms (VQAs) and quantum machine learning (QML) algorithms. qLEET enables the computation of properties such as expressibility and entangling power of a PQC by studying its entanglement spectrum and the distribution of parameterized states produced by it. Furthermore, it allows users to visualize the training trajectories of PQCs along with high-dimensional loss landscapes generated by them for different objective functions. It supports quantum circuits and noise models built using popular quantum computing libraries such as Qiskit, Cirq, and Pyquil. In our work, we demonstrate how qLEET provides opportunities to design and improve hybrid quantum-classical algorithms by utilizing intuitive insights from the ansatz capability and structure of the loss landscape.