AI & ML
Gradient-Based Defensive Poisoning Against System Identification Attacks on Cyber-Physical Systems
arXiv eess.SY0 views
arXiv:2611.24662v1 Announce Type: cross Abstract: AI-driven attacks learn the models of cyber-physical systems (CPS) by passively observing controller and plant signals, then use this knowledge to act covertly. Because covertness depends on the accuracy of the identified model, degrading that accuracy is a viable countermeasure. This work applies Adversarial Machine Learning as a defense, injecting small perturbations into the measurements the attacker observes to degrade an artificial neural network (ANN)-based system identification (SI) attack. The perturbations are derived from input gradients of several independently trained identification models, aggregated into a generalized perturbation, and confined to the transient regimes that carry most of the dynamic information. A sparsity parameter further restricts them to the most influential samples, with negligible cost in effectiveness. On a simulated DC motor networked control system, the proposed perturbations induce larger prediction errors than stochastic poisoning of equal root-mean-square magnitude, and the advantage widens as the magnitude grows. This limits the defense's operating range, which is intended for decoy environments such as CPS honeynets. Additional results indicate that the designed poisoning signals can also degrade N4SID-based system identification, providing evidence of their transferability beyond the ANN-based models originally targeted.
Source:https://arxiv.org/abs/2611.24662PreviousCapacity-Aware Redistribution Control for Mitigation of Cascading Failures in NetworksNextTask-Oriented Network Ranging via Joint Graph and Schedule Design
Related posts
'Insane' to see AI in hands of private giants: Ex-Anthropic researcher to AFP
0Phys.org Tech
Meta says 'reevaluating' 2030 net-zero target amid AI buildout
0Phys.org Tech
Anthropic soars in record Wall Street debut
0Phys.org Tech
Adaptive AI agent connects experimental planning with robotic control in wastewater treatment test
0Phys.org Tech

Black Friday streaming deals live: Apple TV, Paramount+, Crunchyroll, and more have up to 75% off
0TechRadar

It's been the worst week of major outages for 13 years, but don't blame AI
0TechRadar