InfoBot: Transfer and Exploration via the Information BottleneckEI

    Riashat Islam
    Riashat Islam
    Daniel Strouse
    Daniel Strouse
    Zafarali Ahmed
    Zafarali Ahmed
    Cited by: 16|Bibtex|54|

    International Conference on Learning Representations, 2019.

    Abstract:

    A central challenge in reinforcement learning is discovering effective policies for tasks where rewards are sparsely distributed. We postulate that in the absence of useful reward signals, an effective exploration strategy should seek out {it decision states}. These states lie at critical junctions in the state space from where the agent ...More
    Your rating :
    0

     

    Tags
    Comments