Compact quantum kernel-based binary classifier

Abstract Quantum computing opens exciting opportunities for kernel-based machine learning methods, which have broad applications in data analysis. Recent works show that quantum computers can efficiently construct a model of classifier by engineering the interference effect to carry out kernel evaluation parallel. For practical these an important issue is minimize size circuits. We present simplest circuit constructing binary classifier. This achieved generalizing encode labels relative phases state and introducing compact amplitude encoding, encodes two training vectors into one register. When compared known classifier, number qubits reduced steps linearly with respect data. The two-qubit measurement post-selection required previous method simplified single-qubit measurement. Furthermore, final has smaller amount entanglement than method, advocates cost-effectiveness our method. Our design also provides straightforward way handle imbalanced set, often encountered many problems.