Batch-iFDD for Representation Expansion in Large MDPs

@article{Geramifard2013BatchiFDDFR,
  title={Batch-iFDD for Representation Expansion in Large MDPs},
  author={Alborz Geramifard and Thomas J. Walsh and Nicholas Roy and Jonathan P. How},
  journal={CoRR},
  year={2013},
  volume={abs/1309.6831}
}
Conclusion Matching pursuit (MP) techniques are a promising class of feature construction algorithms for value function approximation. Yet until now, applying MP methods required creating a pool of potential features, mandating background knowledge or enumeration of a large feature set, both of which hinder scalability. This paper introduces batch incremental feature dependency discovery (BatchiFDD), an algorithm that is proven to be an MP technique, hence inheriting a provable convergence… CONTINUE READING
Recent Discussions
This paper has been referenced on Twitter 3 times over the past 90 days. VIEW TWEETS

References

Publications referenced by this paper.
Showing 1-10 of 16 references

RLPy: The Reinforcement Learning Library for Education and Research

  • Alborz Geramifard, Robert H Klein, Jonathan P How
  • 2013
1 Excerpt

Similar Papers

Loading similar papers…