Gaussian Imagination in Bandit Learning