qleet 0.1.0.dev0

qLEET is an open-source library for exploring Loss landscape, Expressibility, Entangling capability and Training trajectories of noisy parameterized quantum circuits.

qLEET

qLEET is an open-source library for exploring Loss landscape, Expressibility, Entangling capability and Training trajectories of noisy parameterized quantum circuits.

This project is supported by Unitary Fund.

Features

1. Will support Qiskit’s, Cirq’s and pyQuil's quantum circuits and noise models.
2. Provides opportunities to improve existing algorithms like VQE, QAOA by utilizing intuitive insights from the ansatz capability and structure of loss landscape.
3. Facilitate research in designing new hybrid quantum-classical algorithms.

Examples

Properties of an Ansatz

Ansatz

Expressibility

Solving MAX-CUT using QAOA

Graph

Loss Landscape

Training Path

Contributions

We love your input! We want to make contributing to this project as easy and transparent as possible, whether it's:

• Reporting a bug
• Submitting a fix
• Proposing new features

Feel free to open an issue on this repository or add a pull request to submit your contribution. Adding test cases for any contributions is a requirement for any pull request to be merged

References

1. Expressibility and Entangling Capability of Parameterized Quantum Circuits for Hybrid Quantum‐Classical Algorithms, Sim, S., Johnson, P. D., & Aspuru‐Guzik, A. Advanced Quantum Technologies, 2(12), 1900070. Wiley. (2019)
2. Visualizing the Loss Landscape of Neural Nets, Hao Li, Zheng Xu, Gavin Taylor, Christoph Studer, Tom Goldstein, NIPS 2018, arXiv:1712.09913 [cs.LG] (2018)