GAE: https://arxiv.org/pdf/1506.02438.pdf AC details: http://rail.eecs.berkeley.edu/deeprlcourse/static/slides/lec-6.pdf finite sum optimization for neural network: http://katselis.web.engr.illinois.edu/ECE586/Lecture7.pdf Control Variate: https://arxiv.org/pdf/1611.02247.pdf ACTION-DEPENDENT FACTORIZED BASELINES: https://arxiv.org/pdf/1803.07246.pdf CombineILRL: http://rail.eecs.berkeley.edu/deeprlcourse/static/slides/lec-19.pdf