Gap-Dependent Bounds for Q-Learning using Reference-Advantage Decomposition.

Zhong Zheng, Haochen Zhang, Lingzhou Xue

Browse the full ICLR paper archive.