Resource frugal optimizer for quantum machine learning

Quantum Science and Technology, vol. 8, pp. 045019

Abstract

Abstract Quantum-enhanced data science, also known as quantum machine learning (QML), is of growing interest as an application of near-term quantum computers. Variational QML algorithms have the potential to solve practical problems on real hardware, particularly when involving quantum data. However, training these algorithms can be challenging and calls for tailored optimization procedures. Specifically, QML applications can require a large shot-count overhead due to the large datasets involved. In this work, we advocate for simultaneous random sampling over both the dataset as well as the measurement operators that define the loss function. We consider a highly general loss function that encompasses many QML applications, and we show how to construct an unbiased estimator of its gradient. This allows us to propose a shot-frugal gradient descent optimizer called Refoqus (REsource Frugal Optimizer for QUantum Stochastic gradient descent). Our numerics indicate that Refoqus can save several orders of magnitude in shot cost, even relative to optimizers that sample over measurement operators alone.

Authors 6

1. Charles Moussa corresponding Aachen

   Leiden University

   Affiliation as printed

   Universiteit Leiden LIACS, Niels Bohrweg 1, Leiden, 2333 CA, NETHERLANDS

2. Max Hunter Gordon

   Consejo Superior de Investigaciones Científicas · Instituto de Física Teórica · Universidad Autónoma de Madrid

   Affiliation as printed

   Instituto de Física Teórica, UAM/CSIC, Universidad Autonoma de Madrid, Madrid, Madrid, Madrid, 28049, SPAIN

3. Michał Bączyk

   University of Warsaw

   Affiliation as printed

   University of Warsaw Faculty of Physics, ulica Pasteura 5, Warszawa, 00-681, POLAND

4. Marco Cerezo

   Los Alamos National Laboratory

   Affiliation as printed

   Los Alamos National Laboratory, Los Alamos, Los Alamos, 87545-1663, UNITED STATES

5. Łukasz Cincio

   Los Alamos National Laboratory

   Affiliation as printed

   Theory Division, Los Alamos National Laboratory, P.O. Box 1663, MS B213, Los Alamos, Los Alamos, New Mexico, 87545, UNITED STATES

6. Patrick J. Coles

   Los Alamos National Laboratory

   Affiliation as printed

   Theoretical Division, Los Alamos National Laboratory, Los Alamos, Los Alamos, New Mexico, 87545-1663, UNITED STATES