It is challenging to reduce the complexity of neural networks while maintaining their generalization ability and robustness, especially for practical applications. Conventional solutions for this problem incorporate quantum-inspired neural networks with Kronecker products and hybrid tensor neural networks with MPO factorization and fully-connected layers. Nonetheless, the generalization power and robustness of the fully-connected layers are not as outstanding as circuit models in quantum computing. In this paper, we propose a novel tensor circuit neural network (TCNN) that takes advantage of the characteristics of tensor neural networks and residual circuit models to achieve generalization ability and robustness with low complexity. The proposed activation operation and parallelism of the circuit in complex number field improves its non-linearity and efficiency for feature learning. Moreover, since the feature information exists in the parameters in both the real and imaginary parts in TCNN, an information fusion layer is proposed for merging features stored in those parameters to enhance the generalization capability. Experimental results confirm that TCNN showcases more outstanding generalization and robustness with its average accuracies on various datasets 2\%-3\% higher than those of the state-of-the-art compared models. More significantly, while other models fail to learn features under noise parameter attacking, TCNN still showcases prominent learning capability owing to its ability to prevent gradient explosion. Furthermore, it is comparable to the compared models on the number of trainable parameters and the CPU running time. An ablation study also indicates the advantage of the activation operation, the parallelism architecture and the information fusion layer.
It is challenging to reduce the complexity of neural networks while maintaining their generalization ability and robustness, especially for practical applications. Conventional solutions for this problem incorporate quantum-inspired neural networks with Kronecker products and hybrid tensor neural networks with MPO factorization and fully-connected layers. Nonetheless, the generalization power and robustness of the fully-connected layers are not as outstanding as circuit models in quantum computing. In this paper, we propose a novel tensor circuit neural network (TCNN) that takes advantage of the characteristics of tensor neural networks and residual circuit models to achieve generalization ability and robustness with low complexity. The proposed activation operation and parallelism of the circuit in complex number field improves its non-linearity and efficiency for feature learning. Moreover, since the feature information exists in the parameters in both the real and imaginary parts in TCNN, an information fusion layer is proposed for merging features stored in those parameters to enhance the generalization capability. Experimental results confirm that TCNN showcases more outstanding generalization and robustness with its average accuracies on various datasets 2%-3% higher than those of the state-of-the-art compared models. More significantly, while other models fail to learn features under noise parameter attacking, TCNN still showcases prominent learning capability owing to its ability to prevent gradient explosion. Furthermore, it is comparable to the compared models on the number of trainable parameters and the CPU running time. An ablation study also indicates the advantage of the activation operation, the parallelism architecture and the information fusion layer.
Tensor network, which originates from quantum physics, is emerging as an efficient tool for classical and quantum machine learning. Nevertheless, there still exists a considerable accuracy gap between tensor network and the sophisticated neural network models for classical machine learning. In this work, we combine the ideas of matrix product state (MPS), the simplest tensor network structure, and residual neural network and propose the residual matrix product state (ResMPS). The ResMPS can be treated as a network where its layers map the "hidden" features to the outputs (e.g., classifications), and the variational parameters of the layers are the functions of the features of the samples (e.g., pixels of images). This is different from neural network, where the layers map feed-forwardly the features to the output. The ResMPS can equip with the non-linear activations and dropout layers, and outperforms the state-of-the-art tensor network models in terms of efficiency, stability, and expression power. Besides, ResMPS is interpretable from the perspective of polynomial expansion, where the factorization and exponential machines naturally emerge. Our work contributes to connecting and hybridizing neural and tensor networks, which is crucial to further enhance our understanding of the working mechanisms and improve the performance of both models.
In this paper, we propose two hybrid quantum-inspired neural networks with adaptive residual and dense connections respectively for pattern recognition. We explain the frameworks of the symmetrical circuit models in the quantum-inspired layers in our hybrid models. We also illustrate the potential superiority of our hybrid models to prevent gradient explosion owing to the sine and cosine functions in the quantum-inspired layers. Groups of numerical experiments on generalization power showcase that our hybrid models are comparable to the pure classical models with different noisy datasets utilized. Furthermore, the comparison between our hybrid models and a state-of-the-art hybrid quantum-classical convolutional network demonstrates 3%-4% higher accuracy of our hybrid densely-connected model than the hybrid quantum-classical network. Additionally, compared with other two hybrid quantum-inspired residual networks, our hybrid models showcase a little higher accuracy on image datasets with asymmetrical noises. Simultaneously, in terms of groups of robustness experiments, the outcomes demonstrate that our two hybrid models outperform pure classical models notably in resistance to adversarial parameter attacks with various asymmetrical noises. They also indicate the slight superiority of our densely-connected hybrid model over the hybrid quantum-classical network to both symmetrical and asymmetrical attacks. Meanwhile, the accuracy of our two hybrid models is a little bit higher than that of the two hybrid quantum-inspired residual networks. In addition, an ablation study indicate that the recognition accuracy of our two hybrid models is 2%-3% higher than that of the traditional quantum-inspired neural network without residual or dense connection. Eventually, we discuss the application scenarios of our hybrid models by analyzing their computational complexity.
Driven by the significant advantages offered by quantum computing, research in quantum machine learning has increased in recent years. While quantum speed-up has been demonstrated in some applications of quantum machine learning, a comprehensive understanding of its underlying mechanisms for improved performance remains elusive. Our study address this problem by investigating the functional expressibility of quantum circuits integrated within a convolutional neural network (CNN). Through numerical experiments on the MNIST, Fashion MNIST, and Letter datasets, our hybrid quantum-classical CNN model demonstrates superior feature selection capabilities and substantially reduces the required training steps compared to classical CNNs. Notably, we observe similar performance improvements when incorporating three other quantum-inspired activation functions in classical neural networks, indicating the benefits of adopting quantum-inspired activation functions. Additionally, we developed a hybrid quantum Chebyshev-polynomial network (QCPN) based on the properties of quantum activation functions. We demonstrate that a three-layer QCPN can approximate any continuous function, a feat not achievable by a standard three-layer classical neural network. Our findings suggest that quantum-inspired activation functions can reduce model depth while maintaining high learning capability, making them a promising approach for optimizing large-scale machine-learning models. We also outline future research directions for leveraging quantum advantages in machine learning, aiming to unlock further potential in this rapidly evolving field.
Deep learning has achieved significant success in pattern recognition, with convolutional neural networks (CNNs) serving as a foundational architecture for extracting spatial features from images. Quantum computing provides an alternative computational framework, a hybrid quantum-classical convolutional neural networks (QCCNNs) leverage high-dimensional Hilbert spaces and entanglement to surpass classical CNNs in image classification accuracy under comparable architectures. Despite performance improvements, QCCNNs typically use fixed quantum layers without incorporating trainable quantum parameters. This limits their ability to capture non-linear quantum representations and separates the model from the potential advantages of expressive quantum learning. In this work, we present a hybrid quantum-classical-quantum convolutional neural network (QCQ-CNN) that incorporates a quantum convolutional filter, a shallow classical CNN, and a trainable variational quantum classifier. This architecture aims to enhance the expressivity of decision boundaries in image classification tasks by introducing tunable quantum parameters into the end-to-end learning process. Through a series of small-sample experiments on MNIST, F-MNIST, and MRI tumor datasets, QCQ-CNN demonstrates competitive accuracy and convergence behavior compared to classical and hybrid baselines. We further analyze the effect of ansatz depth and find that moderate-depth quantum circuits can improve learning stability without introducing excessive complexity. Additionally, simulations incorporating depolarizing noise and finite sampling shots suggest that QCQ-CNN maintains a certain degree of robustness under realistic quantum noise conditions. While our results are currently limited to simulations with small-scale quantum circuits, the proposed approach offers a potentially promising direction for hybrid quantum learning in near-term applications.
Image Classification is one of the major tasks in deep learning. Various deep learning methodologies like Quantum Neural Network are playing a vital role in classifying the images. In traditional neural networks neural layers are used to extract the features from the image to train the model where some of the details may be excluded. In quantum methodology parameterised circuits are used to identify complex boundaries to extract high quality features from the images. In the proposed method a hybrid QNN model is developed for binary image classification using quantum computing. Proposed model efficiently works by performing computation on Noise Intermediate-Scale Quantum (NISQ) devices. which represent the current forefront of quantum computing capabilities. This is achieved by integrating a small, two-qubit quantum circuit with a classical convolutional backbone using a tensor product state. The proposed model achieves a classification accuracy of 90.1% on binary image datasets, marking a significant improvement. Furthermore, we evaluate the generalization performance of both the H-QNN and baseline CNN models in image retrieval tasks. The results demonstrate the superior generalization capabilities of the H-QNN, especially in mitigating overfitting on small datasets, making it a promising solution for real-world applications.
We propose a hardware efficient quantum residual neural network which implements residual connections through a deterministic mixture of the identity operation and variational unitaries, enabling fully differentiable training. In contrast to the previous implementation of residual connections, our architecture avoids post-selection while preserving residual learning. Furthermore, we highlight circuit constructions where barren plateaus could be mitigated, which are considered as a major limitation of variational quantum learning models. In order to show the working of our model, we report its application to image classification tasks by training it for MNIST, CIFAR, and SARFish datasets, achieving accuracies of 99\% and 80\% for binary and multi-class classifications, respectively. These accuracies are comparable to previously achieved from the standard variational models, however our model requires 10x fewer gates making it better suited for resource constraint near-term quantum processors. In addition to high accuracies, the proposed architecture also demonstrates adversarial robustness which is another desirable parameter for quantum machine learning models. Overall our architecture offers a new pathway for developing accurate, robust, trainable and hardware efficient quantum machine learning models.