When Critics Disagree: Adaptive Reward Poisoning Attacks in RIS-Aided Wireless Control System

TL;DR

This paper introduces an adaptive reward poisoning attack called DGRP targeting SAC agents in RIS-assisted wireless systems, significantly impairing their performance by exploiting critic disagreements.

Contribution

The paper presents DGRP, a novel disagreement-guided attack method that effectively disrupts DRL-based wireless control systems in RIS environments, highlighting robustness vulnerabilities.

Findings

DGRP significantly reduces RIS's performance benefits.

DGRP causes more damage than baseline attacks.

Attack effectiveness depends on key parameters.

Abstract

Reward-poisoning attacks present a significant risk to learning-based wireless control systems. Given this, we propose a Disagreement-Guided Reward Poisoning (DGRP) adaptive attack on a Soft Actor-Critic (SAC) agent. In a Cognitive Radio Network (CRN) environment assisted by Reconfigurable Intelligent Surfaces (RIS), the SAC agent is tasked with maximizing the long-term secondary users' (SUs) rate by simultaneously optimizing the transmission power of the SU transmitter and the RIS phase shifts. DGRP corrupts rewards, particularly when the SAC dual critics exhibit substantial disagreement-especially in high-leverage, high-uncertainty states-resulting in distorted value estimations and guiding the policy towards suboptimal actions. Our findings demonstrate that DGRP substantially diminishes the performance improvements typically provided by RIS and degrades transmission quality. We further investigate key attack parameters and determine their impact on learning. In comparison to periodic-timing and exploration-triggered baselines, DGRP consistently causes greater damage, highlighting the necessity of considering disagreement-aware threats when evaluating the robustness of Deep Reinforcement Learning (DRL) in RIS-assisted networks.