Security Analysis of Safe and Seldonian Reinforcement Learning Algorithms

Abstract

We analyze the extent to which existing methods rely on accurate training data for a specific class of reinforcement learning (RL) algorithms, known as Safe and Seldonian RL. We introduce a new measure of security to quantify the susceptibility to perturbations in training data by creating an attacker model that represents a worst-case analysis, and show that a couple of Seldonian RL methods are extremely sensitive to even a few data corruptions. We then introduce a new algorithm that is more robust against data corruptions, and demonstrate its usage in practice on some RL problems, including a grid-world and a diabetes treatment simulation.

Cite

Text

Ozisik and Thomas. "Security Analysis of Safe and Seldonian Reinforcement Learning Algorithms." Neural Information Processing Systems, 2020.

Markdown

[Ozisik and Thomas. "Security Analysis of Safe and Seldonian Reinforcement Learning Algorithms." Neural Information Processing Systems, 2020.](https://mlanthology.org/neurips/2020/ozisik2020neurips-security/)

BibTeX

@inproceedings{ozisik2020neurips-security,
  title     = {{Security Analysis of Safe and Seldonian Reinforcement Learning Algorithms}},
  author    = {Ozisik, Pinar and Thomas, Philip S.},
  booktitle = {Neural Information Processing Systems},
  year      = {2020},
  url       = {https://mlanthology.org/neurips/2020/ozisik2020neurips-security/}
}