A Deep Reinforcement Learning Approach to Marginalized Importance
Sampling with the Successor Representation
Scott Fujimoto 1 David Meger 1 Doina Precup 1
Abstract approach can have significantly lower variance than tradi-
tional importance sampling methods (Precup et al., 2001),
Marginalized importance sampling (MIS), which
which consider a product of ratios over traject ...


雷达卡




京公网安备 11010802022788号







