Efficient Bayesian Inference of Sigmoidal Gaussian Cox Processes