Policies for Active Learning from Demonstration.
Lists
Only you see this
Abstract
This paper focuses on Learning from Demonstration, and in particular, on the problem of Active Learning from Demonstration in settings where the amount of data that can be acquired from the demonstrator is limited. In this paper we propose a novel Active Learning from Demonstration approach, SALT, as well as a brief exploration of using a different underlying learning algorithm. We evaluate this approach in two domains: the classic platform game Super Mario Bros. and a grid-based puzzle game called Thermometers.
BibTeX
@inproceedings{packard2017policies,
author = {Brandon Packard and Santiago Ontañón},
title = {{Policies for Active Learning from Demonstration.}},
booktitle = {AAAI Conference on Artificial Intelligence},
year = {2017},
url = {https://openreview.net/forum?id=oosjEj6ijZJ},
}Loading related content...