Manuel Rudolph and Sebastian Schmitt, "Machine Learning on Near-Term Universal Quantum Computers", 1st DPG Fall Meeting - Quantum Science and Information Technologies, 2019.

Implementing near-term quantum computers with a small number of qubits and imperfect gate fidelities for real world challenges has been a flourishing field of research in recent years. Quantum-classical hybrid algorithms with shallow quantum circuits for state preparation are being used with success in fields like quantum chemistry and machine learning. This work focuses on the use of near-term quantum computers for unsupervised machine learning on classical data sets with different model infrastructures. It is shown that the quantum state is able to learn the statistics and correlations of data using shallow variational state preparation. Simple data sets are used to study general aspects such as learning, sampling and generalization of such quantum machine learning implementations in search of practical applications for small quantum machines.