The Virtuous Cycle of Quantum-Classical Machine Learning

Artificial intelligence and quantum computing have both seen tremendous progress in recent years, opening up new avenues for accelerating scientific discovery. The emerging field of quantum machine learning is aimed at utilizing the strengths of both computational paradigms, via the development of learning algorithms or models that benefit from running on quantum devices or from training on quantum data. In this context, classical machine learning has already proven to be useful for addressing challenges in quantum computation; and hardware progress is well underway towards the early fault-tolerant regime. We therefore argue that we are at a pivotal point where both of these paradigms stand to benefit substantially from each other. Here we review the most salient opportunities for machine learning in quantum computing, hoping to establish a virtuous cycle through which both fields can reach their full potential.