Monte Carlo methods are the preferred method of choice when simulating the Boltzmann equation, because of the high-dimensionality of the problem. However, despite their successes, they suffer badly from noise when the strength of the signal is small, as is the case for flows at the micro/nano scale. In this talk I describe existing methodologies for dealing with this problem, and propose a novel method for reducing the noise of stochastic simulations of a Fokker-Planck approximation to the Boltzmann equation.