Targets in Reinforcement Learning to solve Stackelberg Security Games
Machine Learning
2022-12-01 v1 Artificial Intelligence
Computer Science and Game Theory
Multiagent Systems
Machine Learning
Abstract
Reinforcement Learning (RL) algorithms have been successfully applied to real world situations like illegal smuggling, poaching, deforestation, climate change, airport security, etc. These scenarios can be framed as Stackelberg security games (SSGs) where defenders and attackers compete to control target resources. The algorithm's competency is assessed by which agent is controlling the targets. This review investigates modeling of SSGs in RL with a focus on possible improvements of target representations in RL algorithms.
Cite
@article{arxiv.2211.17132,
title = {Targets in Reinforcement Learning to solve Stackelberg Security Games},
author = {Saptarashmi Bandyopadhyay and Chenqi Zhu and Philip Daniel and Joshua Morrison and Ethan Shay and John Dickerson},
journal= {arXiv preprint arXiv:2211.17132},
year = {2022}
}
Comments
Appears in Proceedings of AAAI FSS-22 Symposium "Lessons Learned for Autonomous Assessment of Machine Abilities (LLAAMA)"