S. Li and C. Zhang, “An Optimal Online Method of Selecting Source Policies for Reinforcement Learning”, AAAI, vol. 32, no. 1, Apr. 2018.