Quantum Convolutional Neural Networks: A Survey on Architectures, Applications, and Future Directions

Area of Science:

• Quantum computing
• Artificial intelligence
• Machine learning

Background:

• Quantum convolutional neural networks (QCNNs) integrate quantum principles with deep learning architectures.
• Inspired by classical CNNs, QCNNs leverage quantum phenomena like superposition and entanglement.
• QCNNs have diverse applications, from quantum physics to classical machine learning tasks.

Purpose of the Study:

• To provide a systematic and holistic survey of Quantum Convolutional Neural Networks (QCNNs).
• To offer comparative insights into QCNN architectures, applications, and software toolboxes.
• To identify open challenges and future research directions in the QCNN field.

Main Methods:

• Systematic literature review of QCNN architectures and applications.
• Comparative analysis of different QCNN models (fully quantum, variational, hybrid, graph-based).
• Evaluation of software ecosystems like Qiskit Machine Learning, Pennylane, and TensorFlow Quantum.

Main Results:

• Identified several QCNN architectural variants and their applications.
• Highlighted the role of software development kits in accelerating QCNN research.
• Revealed a lack of unifying taxonomy and comprehensive reviews in the existing literature.

Conclusions:

• QCNNs show significant potential across various scientific and machine learning domains.
• Further research is needed in scalable architectures, fault tolerance, and domain-knowledge integration.
• This survey serves as a foundational reference for advancing QCNN research.