Survey on Hybrid Classical-Quantum Machine Learning Models

Optimizing for a classification problem with a 0–1 loss function is an NP-hard problem [1] even for a simple binary classification task. As Preskill stated “We don't expect a quantum computer to solve worst case instances of NP-hard problems, but it might find better approximate solution or find it faster” [2]. Hybrid Quantum Classical Platforms harness the full capacity of Near-Term Quantum Computers. In this study, a focus is given on the variational circuits based approach accomplishing various machine learning tasks. We will survey a variety of experimental demonstrations conducted on actual quantum hardware and actively developing simulation software, anticipated to have a broad range of real-world applications in this increasingly growing field.