On the Theory of Reinforcement Learning with Once-per-Episode Feedback
Niladri Shekhar Chatterji, Aldo Pacchiano, Peter Bartlett, Michael Jordan
Abstract
We study a theory of reinforcement learning (RL) in which the learner receives binary feedback only once at the end of an episode. While this is an extreme test case for theory, it is also arguably more representative of real-world applications than the traditional requirement in RL practice that the learner receive feedback at every time step. Indeed, in many real-world applications of reinforcement learning, such as self-driving cars and robotics, it is easier to evaluate whether a learner's complete trajectory was either ``good'' or ``bad,'' but harder to provide a reward signal at each step. To show that learning is possible in this more challenging setting, we study the case where trajectory labels are generated by an unknown parametric model, and provide a statistically and computationally efficient algorithm that achieves sublinear regret.
BibTeX
@inproceedings{
chatterji2021on,
title={On the Theory of Reinforcement Learning with Once-per-Episode Feedback},
author={Niladri Shekhar Chatterji and Aldo Pacchiano and Peter Bartlett and Michael Jordan},
booktitle={Advances in Neural Information Processing Systems},
editor={A. Beygelzimer and Y. Dauphin and P. Liang and J. Wortman Vaughan},
year={2021},
url={https://openreview.net/forum?id=c-CD0Yu1ew1}
}