Quantum Convolutional Neural Network: A Hybrid Quantum-Classical Approach for Iris Dataset Classification

Quantum computing is transforming computational paradigms by offering new approaches to solving complex problems, particularly those that push the limits of classical computing. Quantum mechanics' principles, such as superposition, entanglement, and quantum parallelism, allow quantum systems to process information in ways fundamentally distinct from classical systems [1]. These features have the potential to revolutionize areas like machine learning, optimization, and simulation. However, the current limitations of quantum hardware, known as Noisy Intermediate-Scale Quantum (NISQ) devices, prevent the full realization of purely quantum algorithms [2]. In response, hybrid quantum-classical models have emerged as a promising compromise, leveraging the power of quantum computing while maintaining the scalability of classical methods [3]. The concept of Quantum Convolutional Neural Networks (QCNNs), as introduced by Cong et al., further highlights the potential of quantum machine learning, particularly for tasks involving pattern recognition and classification in quantum data [4]. In this study, we introduce an enhanced Quantum Convolutional Neural Network (QCNN) designed to highlight the advantages of hybrid quantum-classical frameworks in machine learning. Our model is applied to the classical Iris dataset, a well-established benchmark in machine learning, which presents a structured yet challenging problem for quantum models.