Density matrix emulation of quantum recurrent neural networks for multivariate time series prediction

Quantum recurrent neural networks (QRNNs) are robust candidates for modelling and predicting future values in multivariate time series. However, the effective implementation of some QRNN models is limited by the need for mid-circuit measurements. Those increase the requirements for quantum hardware,...

Full description

Saved in:
Bibliographic Details
Main Authors: J D Viqueira, D Faílde, M M Juane, A Gómez, D Mera
Format: Article
Language:English
Published: IOP Publishing 2025-01-01
Series:Machine Learning: Science and Technology
Subjects:
Online Access:https://doi.org/10.1088/2632-2153/ad9431
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849770760106672128
author J D Viqueira
D Faílde
M M Juane
A Gómez
D Mera
author_facet J D Viqueira
D Faílde
M M Juane
A Gómez
D Mera
author_sort J D Viqueira
collection DOAJ
description Quantum recurrent neural networks (QRNNs) are robust candidates for modelling and predicting future values in multivariate time series. However, the effective implementation of some QRNN models is limited by the need for mid-circuit measurements. Those increase the requirements for quantum hardware, which in the current noisy intermediate-scale quantum era does not allow reliable computations. Emulation arises as the main near-term alternative to explore the potential of QRNNs, but existing quantum emulators are not dedicated to circuits with multiple intermediate measurements. In this context, we design a specific emulation method that relies on density matrix formalism. Using a compact tensor notation, we provide the mathematical formulation of the operator-sum representation involved. This allows us to show how the present and past information from a time series is transmitted through the circuit, and how to reduce the computational cost in every time step of the emulated network. In addition, we derive the analytical gradient and the Hessian of the network outputs with respect to its trainable parameters, which are needed when the outputs have stochastic noise due to hardware errors and a finite number of circuit shots (sampling). We finally test the presented methods using a hardware-efficient ansatz and four diverse datasets that include univariate and multivariate time series, with and without sampling noise. In addition, we compare the model with other existing quantum and classical approaches. Our results show how QRNNs can be trained with numerical and analytical gradients to make accurate predictions of future values by capturing non-trivial patterns of input series with different complexities.
format Article
id doaj-art-41a5718ee5564d02b9054814d669cbe5
institution DOAJ
issn 2632-2153
language English
publishDate 2025-01-01
publisher IOP Publishing
record_format Article
series Machine Learning: Science and Technology
spelling doaj-art-41a5718ee5564d02b9054814d669cbe52025-08-20T03:02:54ZengIOP PublishingMachine Learning: Science and Technology2632-21532025-01-016101502310.1088/2632-2153/ad9431Density matrix emulation of quantum recurrent neural networks for multivariate time series predictionJ D Viqueira0https://orcid.org/0000-0002-0041-6802D Faílde1https://orcid.org/0000-0002-7685-4331M M Juane2https://orcid.org/0009-0005-0692-7098A Gómez3https://orcid.org/0000-0001-7272-8488D Mera4https://orcid.org/0000-0002-0639-6574Galicia Supercomputing Center (CESGA) , 15705 Santiago de Compostela, Spain; Computer Graphics and Data Engineering (COGRADE), Departamento de Electrónica e Computación, Universidade de Santiago de Compostela , 15782 Santiago de Compostela, SpainGalicia Supercomputing Center (CESGA) , 15705 Santiago de Compostela, SpainGalicia Supercomputing Center (CESGA) , 15705 Santiago de Compostela, SpainGalicia Supercomputing Center (CESGA) , 15705 Santiago de Compostela, SpainComputer Graphics and Data Engineering (COGRADE), Departamento de Electrónica e Computación, Universidade de Santiago de Compostela , 15782 Santiago de Compostela, SpainQuantum recurrent neural networks (QRNNs) are robust candidates for modelling and predicting future values in multivariate time series. However, the effective implementation of some QRNN models is limited by the need for mid-circuit measurements. Those increase the requirements for quantum hardware, which in the current noisy intermediate-scale quantum era does not allow reliable computations. Emulation arises as the main near-term alternative to explore the potential of QRNNs, but existing quantum emulators are not dedicated to circuits with multiple intermediate measurements. In this context, we design a specific emulation method that relies on density matrix formalism. Using a compact tensor notation, we provide the mathematical formulation of the operator-sum representation involved. This allows us to show how the present and past information from a time series is transmitted through the circuit, and how to reduce the computational cost in every time step of the emulated network. In addition, we derive the analytical gradient and the Hessian of the network outputs with respect to its trainable parameters, which are needed when the outputs have stochastic noise due to hardware errors and a finite number of circuit shots (sampling). We finally test the presented methods using a hardware-efficient ansatz and four diverse datasets that include univariate and multivariate time series, with and without sampling noise. In addition, we compare the model with other existing quantum and classical approaches. Our results show how QRNNs can be trained with numerical and analytical gradients to make accurate predictions of future values by capturing non-trivial patterns of input series with different complexities.https://doi.org/10.1088/2632-2153/ad9431quantum computingmachine learningtime series prediction
spellingShingle J D Viqueira
D Faílde
M M Juane
A Gómez
D Mera
Density matrix emulation of quantum recurrent neural networks for multivariate time series prediction
Machine Learning: Science and Technology
quantum computing
machine learning
time series prediction
title Density matrix emulation of quantum recurrent neural networks for multivariate time series prediction
title_full Density matrix emulation of quantum recurrent neural networks for multivariate time series prediction
title_fullStr Density matrix emulation of quantum recurrent neural networks for multivariate time series prediction
title_full_unstemmed Density matrix emulation of quantum recurrent neural networks for multivariate time series prediction
title_short Density matrix emulation of quantum recurrent neural networks for multivariate time series prediction
title_sort density matrix emulation of quantum recurrent neural networks for multivariate time series prediction
topic quantum computing
machine learning
time series prediction
url https://doi.org/10.1088/2632-2153/ad9431
work_keys_str_mv AT jdviqueira densitymatrixemulationofquantumrecurrentneuralnetworksformultivariatetimeseriesprediction
AT dfailde densitymatrixemulationofquantumrecurrentneuralnetworksformultivariatetimeseriesprediction
AT mmjuane densitymatrixemulationofquantumrecurrentneuralnetworksformultivariatetimeseriesprediction
AT agomez densitymatrixemulationofquantumrecurrentneuralnetworksformultivariatetimeseriesprediction
AT dmera densitymatrixemulationofquantumrecurrentneuralnetworksformultivariatetimeseriesprediction