Appendix for Regularized Softmax Deep Multi-Agent Q-Learning Ling Pan

A darker color represents a larger value. Work done while at University of Oxford. B.2 Algorithm and More Details for Computing the Approximate Softmax Operator The full algorithm for computing the approximate softmax operator is in Algorithm 1.Algorithm 1 Approximate softmax operator E.1 Experimental Setup T asks. SC2.4.6.2.69232, and performance is not always comparable across versions. All experiments are run on P100 GPU.