Detection of Man-in-the-Middle Attacks in Model-Free Reinforcement Learning

Rishi Rani, Massimo Franceschetti
Proceedings of The 5th Annual Learning for Dynamics and Control Conference, PMLR 211:993-1007, 2023.

Abstract

This paper proposes a Bellman Deviation algorithm for the detection of man-in-the-middle (MITM) attacks occurring when an agent controls a Markov Decision Process (MDP) system using model-free reinforcement learning. This algorithm is derived by constructing a "Bellman Deviation sequence" and finding stochastic bounds on its running sequence average. We show that an intuitive, necessary and sufficient "informational advantage" condition must be met for the proposed algorithm to guarantee the detection of attacks with high probability, while also avoiding false alarms.

Cite this Paper

BibTeX
@InProceedings{pmlr-v211-rani23a, title = {Detection of Man-in-the-Middle Attacks in Model-Free Reinforcement Learning}, author = {Rani, Rishi and Franceschetti, Massimo}, booktitle = {Proceedings of The 5th Annual Learning for Dynamics and Control Conference}, pages = {993--1007}, year = {2023}, editor = {Matni, Nikolai and Morari, Manfred and Pappas, George J.}, volume = {211}, series = {Proceedings of Machine Learning Research}, month = {15--16 Jun}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v211/rani23a/rani23a.pdf}, url = {https://proceedings.mlr.press/v211/rani23a.html}, abstract = {This paper proposes a Bellman Deviation algorithm for the detection of man-in-the-middle (MITM) attacks occurring when an agent controls a Markov Decision Process (MDP) system using model-free reinforcement learning. This algorithm is derived by constructing a "Bellman Deviation sequence" and finding stochastic bounds on its running sequence average. We show that an intuitive, necessary and sufficient "informational advantage" condition must be met for the proposed algorithm to guarantee the detection of attacks with high probability, while also avoiding false alarms.} }
Endnote
%0 Conference Paper %T Detection of Man-in-the-Middle Attacks in Model-Free Reinforcement Learning %A Rishi Rani %A Massimo Franceschetti %B Proceedings of The 5th Annual Learning for Dynamics and Control Conference %C Proceedings of Machine Learning Research %D 2023 %E Nikolai Matni %E Manfred Morari %E George J. Pappas %F pmlr-v211-rani23a %I PMLR %P 993--1007 %U https://proceedings.mlr.press/v211/rani23a.html %V 211 %X This paper proposes a Bellman Deviation algorithm for the detection of man-in-the-middle (MITM) attacks occurring when an agent controls a Markov Decision Process (MDP) system using model-free reinforcement learning. This algorithm is derived by constructing a "Bellman Deviation sequence" and finding stochastic bounds on its running sequence average. We show that an intuitive, necessary and sufficient "informational advantage" condition must be met for the proposed algorithm to guarantee the detection of attacks with high probability, while also avoiding false alarms.
APA
Rani, R. & Franceschetti, M.. (2023). Detection of Man-in-the-Middle Attacks in Model-Free Reinforcement Learning. Proceedings of The 5th Annual Learning for Dynamics and Control Conference, in Proceedings of Machine Learning Research 211:993-1007 Available from https://proceedings.mlr.press/v211/rani23a.html.

Related Material