Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:2204.01194 · Journal of Quantum Information Science

Continuous Variable Quantum MNIST Classifiers —Classical-Quantum Hybrid Quantum Neural Networks

Sophie Choe, Marek A. Perkowski

Published Jan 1, 2022• 17 citations
arXivHTMLPDF

Abstract

In this paper, classical and continuous variable (CV) quantum neural network hybrid multi-classifiers are presented using the MNIST dataset. Currently available classifiers can classify only up to two classes. The proposed architecture allows networks to classify classes up to nm classes, where n represents cutoff dimension and m the number of qumodes on photonic quantum computers. The combination of cutoff dimension and probability measurement method in the CV model allows a quantum circuit to produce output vectors of size nm. They are then interpreted as one-hot encoded labels, padded with nm - 10 zeros. The total of seven different classifiers is built using 2, 3, …, 6, and 8-qumodes on photonic quantum computing simulators, based on the binary classifier architecture proposed in “Continuous variable quantum neural networks” [1]. They are composed of a classical feed-forward neural network, a quantum data encoding circuit, and a CV quantum neural network circuit. On a truncated MNIST dataset of 600 samples, a 4-qumode hybrid classifier achieves 100% training accuracy.

DOI
10.4236/jqis.2022.122005

Related papers

Hybrid Continuous Variable Quantum Neural Network for Audio Classification

Zainab Aziz, Monika Agrawal

Audio classification is a key application in deep learning, and quantum machine learning (QML) provides new possibilities by capturing complex patterns in signals through parameterized quantum circuits. However, existing qubit-based models struggle to efficiently represent continuous-valued data common in audio processing, limiting their scalability and expressivity. Continuous Variable Quantum Computing (CVQC) offers a natural framework for processing real-valued data such as mel-spectrograms, addressing the limitations of qubit-based systems while introducing nonlinearity through non-gaussian gates to ensure universality. In CVQC, each qumode represents an infinite-dimensional Hilbert space, enabling compact yet expressive information representation. This work introduces a two-qumode Hybrid Continuous Variable Quantum Neural Network (HCVQNN) for audio classification that integrates classical neural layers with a CV quantum circuit for advanced feature extraction and encoding. Experiments conducted using PennyLane demonstrate that HCVQNN achieves comparable audio classification accuracy while requiring fewer resources compared to classical deep networks and qubit-based quantum models.

IEEE International Conference on Acoustics Speech and Signal Processing•Apr 21, 2026•0 citations

Experimentally Realizable Continuous-variable Quantum Neural Networks

Shikha Bangar, George Siopsis, Kübra Yeter Aydeniz

The continuous-variable (CV) quantum computing has shown a great potential for building neural network models. These neural networks can have different levels of quantum- classical hybridization depending on the complexity of the problem. The previous work on CV neural network protocols requires the implementation of non-Gaussian operators in the network. These operators are used to introduce non-linearity which is an essential feature of neural networks. Due to these operators, the protocols become hard to execute experimentally. We built a CV hybrid quantum-classical neural network protocol that can be realized experimentally. Our protocol uses Gaussian gates only with the addition of ancillary qumodes. We achieve non-linearity through measurements on ancillary qumodes. Our gates can be implemented with squeezers and beam splitters; hence our protocol can be realized experimentally. We study two canonical machine learning problems in a supervised learning setting - curve fitting and binary classification to test our neural network. For the curve fitting problem, we used a fully quantum neural network which produced exceptional results as discussed in the paper. The second problem was more complicated in that we found the fraudulent transaction in the credit card data. After training, our model was able to identify genuine data with almost 92% accuracy. After studying both the problem, we can confidently conclude that our protocol gives good results with a great advantage that it can be implemented experimentally with current photonic quantum hardware.

IEEE International Conference on Quantum Computing and Engineering (QCE)•Sep 1, 2022•0 citations

Quantum Transfer Learning for MNIST Classification Using a Hybrid Quantum-Classical Approach

Soumyadip Sarkar

We implement a hybrid quantum-classical model for image classification that compresses MNIST digit images into a low-dimensional feature space and then maps these features onto a 5-qubit quantum state. First, an autoencoder compresses each $28\times28$ image (784 pixels) into a 64-dimensional latent vector, preserving salient features of the digit with minimal reconstruction error. We further reduce the latent representation to 5 principal components using Principal Component Analysis (PCA), to match the 5 available qubits. These 5 features are encoded as rotation angles in a quantum circuit with 5 qubits. The quantum feature map applies single-qubit rotations ($R_y$ gates) proportional to the feature values, followed by a Hadamard gate and a cascade of entangling CNOT gates to produce a non-product entangled state. Measuring the 5-qubit state yields a 32-dimensional probability distribution over basis outcomes, which serves as a quantum-enhanced feature vector for classification. A classical neural network with a softmax output is then trained on these 32-dimensional quantum feature vectors to predict the digit class. We evaluate the hybrid model on the MNIST dataset and compare it to a purely classical baseline that uses the 64-dimensional autoencoder latent features for classification. The results show that the hybrid model can successfully classify digits, demonstrating the feasibility of integrating quantum computing in the classification pipeline, although its accuracy (about 75\% on test data) currently falls below the classical baseline (about 98\% on the same compressed data).

arXiv:2408.03351•Aug 5, 2024•6 citations

Quantum-Classical Neural Networks: A Hybrid Approach to MNIST Classification

Tur Mohammad Mubtasim, Md. Siam Uddin Molla Antor, Sheikh Shifat

The combination of the principles of quantum computing and classical deep learning is a potentially fruitful direction to develop the field of artificial intelligence, but the benchmarking of such hybrid architectures against standard computer vision problems is an under-explored area. The paper explores the behavior of Hybrid Quantum-Classical Neural Network (HQNN) in the classification of images into multiple classes using the MNIST database. Our architecture is an HQNN implementing an 8-qubit and 2-layer variational circuit, and it is parameterized by a classical multi-layer perceptron (MLP). To make the evaluation process rigorous and fair, the hybrid was benchmarked on a purely classical MLP, which had almost the same parameter footprint ( 105,000 parameters). Results over 50 training epochs show that the classical MLP reached its highest test accuracy of 96.50% and Macro F1 of 0.96. The HQNN was able to converge and learn the data of high dimensions delivering a competitive but slightly lower maximum accuracy of 94.35% and Macro F1 of 0.94. Moreover, there was a high computational cost on the simulation of the quantum layers, which increased the training time by 233 times the classical baseline. These results suggest that while the quantum layer is expressive and capable of learning complex representations, the current state of quantum simulation on classical hardware is a significant bottleneck. Which can be reduced by QPU/NISQ hardware, lighter quantum circuits and better dta encoding.

International Conference on Power, Electronics, Communications, Computing, and Intelligent Infrastructure (PECCII)•Jun 1, 2026•0 citations

Continuous-variable quantum neural networks

Nathan Killoran, Thomas R. Bromley, Juan Miguel Arrazola +3 more

We introduce a general method for building neural networks on quantum computers. The quantum neural network is a variational quantum circuit built in the continuous-variable (CV) architecture, which encodes quantum information in continuous degrees of freedom such as the amplitudes of the electromagnetic field. This circuit contains a layered structure of continuously parameterized gates which is universal for CV quantum computation. Affine transformations and nonlinear activation functions, two key elements in neural networks, are enacted in the quantum network using Gaussian and non-Gaussian gates, respectively. The non-Gaussian gates provide both the nonlinearity and the universality of the model. Due to the structure of the CV model, the CV quantum neural network can encode highly nonlinear transformations while remaining completely unitary. We show how a classical network can be embedded into the quantum formalism and propose quantum versions of various specialized models such as convolutional, recurrent, and residual networks. Finally, we present numerous modeling experiments built with the strawberry fields software library. These experiments, including a classifier for fraud detection, a network which generates tetris images, and a hybrid classical-quantum autoencoder, demonstrate the capability and adaptability of CV quantum neural networks.

•Aug 1, 2018•358 citations

A Project Submitted to the Graduate Faculty of the State University of New York Polytechnic Institute in Partial Fulfillment of the Requirements for the Degree of Master of Science Utica, New York

Hussam Muhammad Anwar

This project compares classical, quantum, and hybrid neural network architectures for multiclass classification on the MNIST and CIFAR10 datasets. The models include a classical convolutional network, two pure quantum neural networks constructed with TorchQuantum, a quanvolutional model that applied fixed random quantum filters to image patches, and hybrid quantum classical networks that combine classical convolution with trainable quantum layers. In addition, a quantum classical quantum convolutional neural network (QCQ-CNN) inspired by recent works is implemented, as well as a separate Qiskit based QCQ-CNN that uses the ZZFeatureMap for data encoding and the RealAmplitudes ansatz for variational processing. Most models were trained under shared hyperparameters and full multiclass supervision, while the Qiskit based QCQ-CNN was trained on a reduced MNIST and CIFAR10 subset in a One-vs-All setting due to the computational cost of simulating quantum circuits. Across both datasets, the classical convolutional neural network achieved the highest accuracy. The pure quantum models performed well and showed small gaps between training and testing accuracy, which suggests that the models did not overfit and were stable. The quanvolutional model produced expressive features but heavily overfit. The TorchQuantum QCQ-CNN achieved lower accuracy than the classical baseline but showed a consistent generalization and avoided the overfitting seen in the quanvolutional mode. The One-vs-All v QCQ-CNN had a strong accuracy for each class, showing that hybrid neural networks can perform well when the task is reduced into smaller subproblems. These results highlight the strengths and weaknesses of current quantum and hybrid neural networks in practical scenarios. The experiments illustrate that shallow quantum circuits can be integrated into classical neural networks to assist in their generalization ability even while the quantum aspects are simulated.

•Dec 11, 2025•0 citations
In this paper, classical and continuous variable (CV) quantum neural network hybrid multi-classifiers are presented using the MNIST dataset. Currently available classifiers can classify only up to two classes. The proposed architecture allows networks to classify classes up to nm classes, where n represents cutoff dimension and m the number of qumodes on photonic quantum computers. The combination of cutoff dimension and probability measurement method in the CV model allows a quantum circuit to produce output vectors of size nm. They are then interpreted as one-hot encoded labels, padded with nm - 10 zeros. The total of seven different classifiers is built using 2, 3, …, 6, and 8-qumodes on photonic quantum computing simulators, based on the binary classifier architecture proposed in “Continuous variable quantum neural networks” [1]. They are composed of a classical feed-forward neural network, a quantum data encoding circuit, and a CV quantum neural network circuit. On a truncated MNIST dataset of 600 samples, a 4-qumode hybrid classifier achieves 100% training accuracy.
Audio classification is a key application in deep learning, and quantum machine learning (QML) provides new possibilities by capturing complex patterns in signals through parameterized quantum circuits. However, existing qubit-based models struggle to efficiently represent continuous-valued data common in audio processing, limiting their scalability and expressivity. Continuous Variable Quantum Computing (CVQC) offers a natural framework for processing real-valued data such as mel-spectrograms, addressing the limitations of qubit-based systems while introducing nonlinearity through non-gaussian gates to ensure universality. In CVQC, each qumode represents an infinite-dimensional Hilbert space, enabling compact yet expressive information representation. This work introduces a two-qumode Hybrid Continuous Variable Quantum Neural Network (HCVQNN) for audio classification that integrates classical neural layers with a CV quantum circuit for advanced feature extraction and encoding. Experiments conducted using PennyLane demonstrate that HCVQNN achieves comparable audio classification accuracy while requiring fewer resources compared to classical deep networks and qubit-based quantum models.
The continuous-variable (CV) quantum computing has shown a great potential for building neural network models. These neural networks can have different levels of quantum- classical hybridization depending on the complexity of the problem. The previous work on CV neural network protocols requires the implementation of non-Gaussian operators in the network. These operators are used to introduce non-linearity which is an essential feature of neural networks. Due to these operators, the protocols become hard to execute experimentally. We built a CV hybrid quantum-classical neural network protocol that can be realized experimentally. Our protocol uses Gaussian gates only with the addition of ancillary qumodes. We achieve non-linearity through measurements on ancillary qumodes. Our gates can be implemented with squeezers and beam splitters; hence our protocol can be realized experimentally. We study two canonical machine learning problems in a supervised learning setting - curve fitting and binary classification to test our neural network. For the curve fitting problem, we used a fully quantum neural network which produced exceptional results as discussed in the paper. The second problem was more complicated in that we found the fraudulent transaction in the credit card data. After training, our model was able to identify genuine data with almost 92% accuracy. After studying both the problem, we can confidently conclude that our protocol gives good results with a great advantage that it can be implemented experimentally with current photonic quantum hardware.
We implement a hybrid quantum-classical model for image classification that compresses MNIST digit images into a low-dimensional feature space and then maps these features onto a 5-qubit quantum state. First, an autoencoder compresses each 28×2828\times2828×28 image (784 pixels) into a 64-dimensional latent vector, preserving salient features of the digit with minimal reconstruction error. We further reduce the latent representation to 5 principal components using Principal Component Analysis (PCA), to match the 5 available qubits. These 5 features are encoded as rotation angles in a quantum circuit with 5 qubits. The quantum feature map applies single-qubit rotations (RyR_yRy​ gates) proportional to the feature values, followed by a Hadamard gate and a cascade of entangling CNOT gates to produce a non-product entangled state. Measuring the 5-qubit state yields a 32-dimensional probability distribution over basis outcomes, which serves as a quantum-enhanced feature vector for classification. A classical neural network with a softmax output is then trained on these 32-dimensional quantum feature vectors to predict the digit class. We evaluate the hybrid model on the MNIST dataset and compare it to a purely classical baseline that uses the 64-dimensional autoencoder latent features for classification. The results show that the hybrid model can successfully classify digits, demonstrating the feasibility of integrating quantum computing in the classification pipeline, although its accuracy (about 75\% on test data) currently falls below the classical baseline (about 98\% on the same compressed data).
The combination of the principles of quantum computing and classical deep learning is a potentially fruitful direction to develop the field of artificial intelligence, but the benchmarking of such hybrid architectures against standard computer vision problems is an under-explored area. The paper explores the behavior of Hybrid Quantum-Classical Neural Network (HQNN) in the classification of images into multiple classes using the MNIST database. Our architecture is an HQNN implementing an 8-qubit and 2-layer variational circuit, and it is parameterized by a classical multi-layer perceptron (MLP). To make the evaluation process rigorous and fair, the hybrid was benchmarked on a purely classical MLP, which had almost the same parameter footprint ( 105,000 parameters). Results over 50 training epochs show that the classical MLP reached its highest test accuracy of 96.50% and Macro F1 of 0.96. The HQNN was able to converge and learn the data of high dimensions delivering a competitive but slightly lower maximum accuracy of 94.35% and Macro F1 of 0.94. Moreover, there was a high computational cost on the simulation of the quantum layers, which increased the training time by 233 times the classical baseline. These results suggest that while the quantum layer is expressive and capable of learning complex representations, the current state of quantum simulation on classical hardware is a significant bottleneck. Which can be reduced by QPU/NISQ hardware, lighter quantum circuits and better dta encoding.
We introduce a general method for building neural networks on quantum computers. The quantum neural network is a variational quantum circuit built in the continuous-variable (CV) architecture, which encodes quantum information in continuous degrees of freedom such as the amplitudes of the electromagnetic field. This circuit contains a layered structure of continuously parameterized gates which is universal for CV quantum computation. Affine transformations and nonlinear activation functions, two key elements in neural networks, are enacted in the quantum network using Gaussian and non-Gaussian gates, respectively. The non-Gaussian gates provide both the nonlinearity and the universality of the model. Due to the structure of the CV model, the CV quantum neural network can encode highly nonlinear transformations while remaining completely unitary. We show how a classical network can be embedded into the quantum formalism and propose quantum versions of various specialized models such as convolutional, recurrent, and residual networks. Finally, we present numerous modeling experiments built with the strawberry fields software library. These experiments, including a classifier for fraud detection, a network which generates tetris images, and a hybrid classical-quantum autoencoder, demonstrate the capability and adaptability of CV quantum neural networks.
This project compares classical, quantum, and hybrid neural network architectures for multiclass classification on the MNIST and CIFAR10 datasets. The models include a classical convolutional network, two pure quantum neural networks constructed with TorchQuantum, a quanvolutional model that applied fixed random quantum filters to image patches, and hybrid quantum classical networks that combine classical convolution with trainable quantum layers. In addition, a quantum classical quantum convolutional neural network (QCQ-CNN) inspired by recent works is implemented, as well as a separate Qiskit based QCQ-CNN that uses the ZZFeatureMap for data encoding and the RealAmplitudes ansatz for variational processing. Most models were trained under shared hyperparameters and full multiclass supervision, while the Qiskit based QCQ-CNN was trained on a reduced MNIST and CIFAR10 subset in a One-vs-All setting due to the computational cost of simulating quantum circuits. Across both datasets, the classical convolutional neural network achieved the highest accuracy. The pure quantum models performed well and showed small gaps between training and testing accuracy, which suggests that the models did not overfit and were stable. The quanvolutional model produced expressive features but heavily overfit. The TorchQuantum QCQ-CNN achieved lower accuracy than the classical baseline but showed a consistent generalization and avoided the overfitting seen in the quanvolutional mode. The One-vs-All v QCQ-CNN had a strong accuracy for each class, showing that hybrid neural networks can perform well when the task is reduced into smaller subproblems. These results highlight the strengths and weaknesses of current quantum and hybrid neural networks in practical scenarios. The experiments illustrate that shallow quantum circuits can be integrated into classical neural networks to assist in their generalization ability even while the quantum aspects are simulated.