Corpus ID: 209376316

Planning with Abstract Learned Models While Learning Transferable Subtasks

@article{Winder2020PlanningWA,
  title={Planning with Abstract Learned Models While Learning Transferable Subtasks},
  author={John Winder and Stephanie Milani and Matthew Landen and Erebus Oh and Shane Parr and Shawn Squire and Marie desJardins and Cynthia Matuszek},
  journal={ArXiv},
  year={2020},
  volume={abs/1912.07544}
}
  • John Winder, Stephanie Milani, +5 authors Cynthia Matuszek
  • Published in AAAI 2020
  • Mathematics, Computer Science
  • ArXiv
  • We introduce an algorithm for model-based hierarchical reinforcement learning to acquire self-contained transition and reward models suitable for probabilistic planning at multiple levels of abstraction. We call this framework Planning with Abstract Learned Models (PALM). By representing subtasks symbolically using a new formal structure, the lifted abstract Markov decision process (L-AMDP), PALM learns models that are independent and modular. Through our experiments, we show how PALM… CONTINUE READING

    Create an AI-powered research feed to stay up to date with new papers like this posted to ArXiv

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 41 REFERENCES

    Data-Efficient Hierarchical Reinforcement Learning

    VIEW 5 EXCERPTS
    HIGHLY INFLUENTIAL

    Hierarchical model-based reinforcement learning: R-max + MAXQ

    VIEW 3 EXCERPTS
    HIGHLY INFLUENTIAL

    Hierarchical Reinforcement Learning with the MAXQ Value Function Decomposition

    VIEW 5 EXCERPTS
    HIGHLY INFLUENTIAL

    An investigation of model-free planning

    VIEW 1 EXCERPT