PATROL: Provable Defense against Adversarial Policy in Two-player Games.

Wenbo Guo Xian Wu Lun Wang Xinyu Xing Dawn Song

Published in: USENIX Security Symposium (2023)

Keyphrases

two player games
evaluation function
optimal policy
game tree
multi agent
data structure
machine learning
temporal difference learning
games played
reinforcement learning
learning process
dynamic programming
sufficient conditions
orders of magnitude
statistically significant
game playing