Quantum-Enhanced Support Vector Regression for Full-Duplex Self-Interference Cancellation

This letter proposes a quantum-enhanced support vector regression (QSVR) approach to cancel the self-interference in full-duplex transceivers. The proposed approach leverages quantum feature maps, incorporating quantum circuits, to encode classical data into a higher-dimensional feature space. Simulation results unveil a substantial performance enhancement by the proposed QSVR approach over the classical SVR and state-of-the-art hybrid quantum-classical neural network in the literature when being trained with a smaller set of training data.