We propose to leverage epistemic uncertainty about constraint satisfaction of a reinforcement learner in safety critical domains. We introduce a framework for specification of requirements for reinforcement learners in constrained settings, including confidence about results. We show that an agent's confidence in constraint satisfaction provides a useful signal for balancing optimization and safety in the learning process.
@article{arxiv.2005.03898,
title = {Synthesizing Safe Policies under Probabilistic Constraints with Reinforcement Learning and Bayesian Model Checking},
author = {Lenz Belzner and Martin Wirsing},
journal= {arXiv preprint arXiv:2005.03898},
year = {2021}
}