Projective simulation with optical frequency combs: a continuous-variable approach on graph states
The aim of this Discovery project is to develop the theory of a quantum-enhanced reinforcement learning agent that is implementable using quantum optical frequency combs. To this end, we will bring together recent theoretical and experimental achievements within a highly multidisciplinary framework. First, we will recast the decision-making process of the agent into a picture that draws upon the literature of graph states and measurement-based quantum computation. Then, we will conform this agent to a quantum optical platform that is capable to support the above theoretical tools. This project holds the potential to enhance the efficiency of reinforcement learning agents run on classical processors, as well as to enhance quantum optical technologies by equipping them with an embedded artificial intelligence.