Quantum learning speedup in binary classification

We compare quantum and classical machines designed to learn a binary classification problem in order to address how quantum system improves the machine learning behaviour. Two machines consist of the same number of operations and control parameters, but only quantum machine plays with quantum coherence naturally induced by unitary operators. We show that quantum superposition enables quantum learning faster than classical learning by expanding the solution regions. This is also demonstrated by numerical simulations with a standard feedback model, random search, and a practical model, differential evolution.