Preprint
Aug 2026
Robustness of Anomaly Detection Models for Industrial Control Systems under Training-Time Data Contamination
The findings highlight the importance of training-data integrity in ML-enabled ICS monitoring, subject to the evaluated dataset, models, and threat assumptions, and show that robustness is strongly model-dependent and cannot be predicted from clean-data performance alone.
M. Ozbek, Taiwo P. Ojo, Pooria Madani et al.
· 0 citations