Quickest Change Detection in Discrete-Time in Presence of a Covert Adversary

TL;DR

This paper analyzes the limits of quickest change detection in discrete time when facing a covert adversary, providing asymptotic characterizations of detection delay and false alarm time under adversarial conditions.

Contribution

It extends classical change detection theory by incorporating an adversary who chooses post-change distributions to remain covert, deriving asymptotic expressions for detection delay and false alarm time.

Findings

Asymptotic expressions for average detection delay (ADD) and false alarm time (AT2FA) under adversarial conditions.

Identification of critical scaling laws for covert behavior, with ADD scaling linearly with false alarm constraint parameter.

Explicit characterization of ADD for Gaussian and Exponential models as a function of KL divergence and false alarm parameter.

Abstract

We study the problem of covert quickest change detection in a discrete-time setting, where a sequence of observations undergoes a distributional change at an unknown time. Unlike classical formulations, we consider a covert adversary who has knowledge of the detector's false alarm constraint parameter $\gamma$ and selects a stationary post-change distribution that depends on it, seeking to remain undetected for as long as possible. Building on the theoretical foundations of the CuSum procedure, we rigorously characterize the asymptotic behavior of the average detection delay (ADD) and the average time to false alarm (AT2FA) when the post-change distribution converges to the pre-change distribution as $\gamma \to \infty$. Our analysis establishes exact asymptotic expressions for these quantities, extending and refining classical results that no longer hold in this regime. We identify the critical scaling laws governing covert behavior and derive explicit conditions under which an adversary can maintain covertness, defined by ADD = $\Theta(\gamma)$, whereas in the classical setting, ADD grows only as $\mathcal{O}(\log \gamma)$. In particular, for Gaussian and Exponential models under adversarial perturbations of their respective parameters, we asymptotically characterize ADD as a function of the Kullback--Leibler divergence between the pre- and post-change distributions and $\gamma$.