On Modifying the Variational Quantum Singular Value Decomposition Algorithm

TL;DR

This paper proposes two modifications to a variational quantum SVD algorithm, including an improved objective function and a new method for computing expectation values, leading to enhanced performance.

Contribution

It introduces a novel objective function and a new expectation value computation method for the variational quantum SVD algorithm, with benchmarking results.

Findings

Modified objective function shows improved algorithm performance

New expectation value method enhances computational efficiency

Benchmarking confirms better accuracy and speed

Abstract

In this work, we discuss two modifications that can be made to a known variational quantum singular value decomposition algorithm popular in the literature. The first is a change to the objective function which hints at improved performance of the algorithm. The second modification introduces a new way of computing expectation values of general matrices, which is a key step in the algorithm. We then benchmark this modified algorithm and compare the performance of our new objective function with the existing one.