The curse of dimensionality is ubiquitous in machine learning (ML) modeling, stochastic control and reinforcement learning, arising in a probabilistic sense, with strong connections to quantum mechanics.
Typical approaches in ML and controls utilize local approximations, since the globally optimal solution requires solving a problem which scales exponentially with the number of data points or dimensionality of the problem.
Hypothetically, is it possible to use the fundamental connections between these phenomena (whether noisy data or stochastic processes, including quantum phenomena) to solve the problem (rather than re-fashioning classical algorithms to solve NP hard problems) via quantum computing?
