Repeated Inverse Reinforcement Learning
Artificial Intelligence
2017-11-07 v3 Machine Learning
Abstract
We introduce a novel repeated Inverse Reinforcement Learning problem: the agent has to act on behalf of a human in a sequence of tasks and wishes to minimize the number of tasks that it surprises the human by acting suboptimally with respect to how the human would have acted. Each time the human is surprised, the agent is provided a demonstration of the desired behavior by the human. We formalize this problem, including how the sequence of tasks is chosen, in a few different ways and provide some foundational results.
Cite
@article{arxiv.1705.05427,
title = {Repeated Inverse Reinforcement Learning},
author = {Kareem Amin and Nan Jiang and Satinder Singh},
journal= {arXiv preprint arXiv:1705.05427},
year = {2017}
}
Comments
The first two authors contributed equally to this work. The paper appears in NIPS 2017