Ans\"atze for Noisy Variational Quantum Eigensolvers

TL;DR

This paper analyzes various ansätze for noisy Variational Quantum Eigensolvers, focusing on their noise resilience and effectiveness in quantum chemistry applications on current quantum hardware.

Contribution

It provides a comparative analysis of different ansätze, especially dynamic ADAPT-VQEs, highlighting how pool selection and criteria affect performance under noise.

Findings

Dynamic ansätze show improved noise resilience.

Pool choice significantly impacts ansatz efficiency.

Certain ansätze outperform traditional methods in noisy conditions.

Abstract

The hardware requirements of useful quantum algorithms remain unmet by the quantum computers available today. Because it was designed to soften these requirements, the Variational Quantum Eigensolver (VQE) has gained popularity as a contender for a chance at quantum advantage with near-term quantum computers. The ansatz, a parameterized circuit that prepares trial states, can dictate the success (or lack thereof) of a VQE. Too deep ans\"atze can hinder near-term viability, or lead to trainability issues that render the algorithm inefficient. The purpose of this thesis was to analyse different ans\"atze proposed for quantum chemistry, examining their noise-resilience and viability in state-of-the-art quantum computers. In particular, dynamic ans\"atze (ADAPT-VQEs) were explored, and the impact of the choice of pool and selection criterion on their performance was analysed.