Our paper on policy gradient for non-rectangular robust MDPs is accepted to SIAM Journal on Optimization.