A Tutorial on Optimal Control and Reinforcement Learning methods for Quantum Technologies

Abstract

Quantum Optimal Control is an established field of research which is necessary for the development of Quantum Technologies. Quantum Machine Learning is a fast emerging field in which the theory of Quantum Mechanics and Machine Learning fuse together in order to learn and benefit from each other. In particular, Reinforcement Learning has a direct application in control problems. In this tutorial we introduce the methods of Quantum Optimal Control and Reinforcement Learning by applying them to the problem of three-level population transfer. The jupyter notebooks to reproduce some of our results are open-sourced and available on github 1 .

Optimized state transfer in systems of ultrastrongly coupled matter and radiation

“…In order to improve on the results of [20] in what follows we investigate the optimal timedependence of g i (t) using the Quantum Optimal Control (QOC) [14,1] tools developed in [13].…”

“…We found optimized g opt i (t) yielding larger efficiency with a time duration shorter than for Gaussian pulses with the same maximal strength g 0 . It is likely that QOC, machine learning techniques [13,4] or superadiabatic driving [12] could further improve the transfer efficiency of such operations exploiting also modulation of detunings [8,6,7].…”

One decade of quantum optimal control in the chopped random basis

“…Indeed, the best choice of the maximization algorithm is most likely problem-dependent and many examples of such possible choices have been given in section 2. Machine learning tools could be used to systematically utilize the acquired experience for a given new QOC problem [251]. The same arguments can be made for the choice of the basis functions.…”