Discretization-free Knowledge Gradient Methods for Bayesian Optimization | Read Paper on Bytez