Driving Atoms with Light of Arbitrary Statistics

Since it is now possible to produce reliable sources of squeezed, antibunched and sub-Poissonian light, it is relevant to consider what effects these might have when impinged on quantum systems. Two questions immediately arise: Given a source of non-classical light, how does one work out the effect of coupling it to a quantum system? In a theoretical context, how does one specify a non-classical light beam? With thermal or squeezed light beams this is normally done by assuming that they are Gaussian, which is theoretically true, though in principle practical squeezed beams may deviate from being exactly Gaussian. For Gaussian beams the means and second order correlation functions specify all higher correlation functions, so that one need only specify these to specify the beam exactly. There is no such simple method for non-Gaussian beams, such as antibunched beams.