Gaussian-optimized preparation of non-Gaussian pure states

Non-Gaussian states are highly sought-after resources in continuous-variable quantum optical information processing protocols. We outline a method for the optimized preparation of any pure non-Gaussian state to a given desired accuracy. Our proposal arises from two connected concepts. First, we define the operational cost of a desired state as the largest Fock state required for its approximate preparation. Second, we suggest that this non-Gaussian operational cost can be reduced by judicial application of optimized Gaussian operations. In particular, we identify a minimal core non-Gaussian state for any target pure state, which is related to the core state by Gaussian operations alone. We demonstrate this method for Schrodinger cat states.