K. Lee and K.-E. Kim, “Tighter Value Function Bounds for Bayesian Reinforcement Learning”, AAAI, vol. 29, no. 1, Mar. 2015.