MicroCloud Hologram Inc. Releases Hybrid Quantum-Classical Convolutional Neural Network, Achieving New Breakthrough in MNIST Multi-Class Classification

Core Insights - MicroCloud Hologram Inc. has proposed a Quantum Convolutional Neural Network (QCNN) that utilizes hybrid quantum-classical learning, achieving accuracy comparable to classical Convolutional Neural Networks (CNNs) on the MNIST dataset, indicating the practical feasibility of quantum computing in machine learning tasks [1][3][12] Company Overview - MicroCloud Hologram Inc. focuses on providing advanced holographic technology services, including high-precision holographic LiDAR solutions and holographic digital twin technology [15] - The company plans to invest over $400 million in cutting-edge technology sectors, including quantum computing, quantum holography, and artificial intelligence [15] Technological Innovation - The proposed QCNN leverages a hybrid quantum-classical learning framework, combining classical optimizers with quantum circuits to enhance feature extraction and classification tasks [4][11] - The architecture includes a novel Quantum Perceptron model that improves feature extraction and nonlinear mapping capabilities [11] - The quantum circuit employs eight qubits for data encoding and four auxiliary qubits to enhance expressive power, effectively mapping input data for classification [5][11] Application Potential - The QCNN can be applied to various complex datasets and tasks, such as multi-class traffic sign recognition in autonomous driving and multi-class classification of lesions in medical imaging [12][13] - The integration of quantum computing with classical learning offers significant advantages in energy efficiency, parameter efficiency, and computational acceleration for enterprises [13]