Principled Exploration via Optimistic Bootstrapping and Backward Induction

    Chenjia Bai 1 Lingxiao Wang 2 Lei Han 3 Jianye Hao 4 Animesh Garg 5 Peng Liu 1 Zhaoran Wang 2

             Abstract                  2007; Jin et al., 2018) is a principled approach for efficient
    One principled approach for provably efficient        exploration with well theoretical guarantees. In tabular
    exploration is incorporating the upper confidence       cases, the optimism-based methods incorporate th ...