Generalizing Apprenticeship Learning Across Hypothesis Classes
Abstract
This paper develops a generalized apprenticeship learning protocol for reinforcement-learning agents with access to a teacher who provides policy traces (transition and reward observations). We characterize sufficient conditions of the underlying models for efficient apprenticeship learning and link this criteria to two established learn ability classes (KWIK and Mistake Bound). We then construct efficient apprenticeship-learning algorithms in a number of domains, including two types of relational MDPs. We instantiate our approach in a software agent and a robot agent that learn effectively from a human teacher.
Cite
Text
Walsh et al. "Generalizing Apprenticeship Learning Across Hypothesis Classes." International Conference on Machine Learning, 2010.Markdown
[Walsh et al. "Generalizing Apprenticeship Learning Across Hypothesis Classes." International Conference on Machine Learning, 2010.](https://mlanthology.org/icml/2010/walsh2010icml-generalizing/)BibTeX
@inproceedings{walsh2010icml-generalizing,
title = {{Generalizing Apprenticeship Learning Across Hypothesis Classes}},
author = {Walsh, Thomas J. and Subramanian, Kaushik and Littman, Michael L. and Diuk, Carlos},
booktitle = {International Conference on Machine Learning},
year = {2010},
pages = {1119-1126},
url = {https://mlanthology.org/icml/2010/walsh2010icml-generalizing/}
}