In this paper, we identify and study a fundamental, yet underexplored, phenomenon in security games, which we term the Curse of Correlation (CoC). Specifically, we observe that there is inevitable correlation among the protection status of different targets. Such correlation is a crucial concern, especially in spatio-temporal domains like conservation area patrolling, where attackers can monitor patrollers at certain areas and then infer their patrolling routes using such correlation. To mitigate this issue, we introduce the principle of max-entropy to security games, and focus on designing entropy-maximizing defending strategies for the spatio-temporal security game -- a major victim of CoC. We prove that the problem is #P-hard in general, but propose efficient algorithms in well-motivated special settings. Our experiments show significant advantages of the max-entropy algorithms against previous algorithms.
from cs.AI updates on arXiv.org http://ift.tt/2ml7hd0
via IFTTT
No comments:
Post a Comment