Entropy-based Training Methods for Scalable Neural Implicit Samplers | Read Paper on Bytez