Quantum Neural Network Realization of XOR on a Desktop Quantum Computer

Quantum neural networks leverage quantum computing to address machine-learning problems beyond the capabilities of classical computing. In this study, we demonstrate a quantum neural network that learns the nonlinear XOR function on a desktop quantum computer. The XOR task is a nonlinear benchmark that cannot be solved by a single-layer perceptron, making it an excellent test for quantum machine learning. We trained a variational quantum circuit model in a simulation using the PennyLane framework to learn the two-bit XOR mapping. After obtaining the circuit parameters in the simulation, the trained quantum neural network was deployed on a two-qubit Nuclear Magnetic Resonance-based desktop quantum computer operating at room temperature to evaluate the actual hardware performance. The experimental quantum state fidelity reached approximately 98.85% (Ry) and 99.35% Rx, and the overall average purity was 95.16% Ry and 97.43% Rx, indicating excellent agreement between the expected and measured results. These positive outcomes underscore the feasibility of quantum machine learning on small-scale quantum hardware, marking an important step toward practical quantum artificial intelligence applications.