Resilient Machine Learning for Networked Cyber Physical Systems: A Survey for Machine Learning Security to Securing Machine Learning for CPS
The interactions between resilient CPS using ML and resilient ML when applied in CPS are surveyed to have a thorough understanding of recent advances on ML-based security and securing ML for CPS and countermeasures, as well as research trends in this active research area.
Abstract
Cyber Physical Systems (CPS) are characterized by their ability to integrate\nthe physical and information or cyber worlds. Their deployment in critical\ninfrastructure have demonstrated a potential to transform the world. However,\nharnessing this potential is limited by their critical nature and the far\nreaching effects of cyber attacks on human, infrastructure and the environment.\nAn attraction for cyber concerns in CPS rises from the process of sending\ninformation from sensors to actuators over the wireless communication medium,\nthereby widening the attack surface. Traditionally, CPS security has been\ninvestigated from the perspective of preventing intruders from gaining access\nto the system using cryptography and other access control techniques. Most\nresearch work have therefore focused on the detection of attacks in CPS.\nHowever, in a world of increasing adversaries, it is becoming more difficult to\ntotally prevent CPS from adversarial attacks, hence the need to focus on making\nCPS resilient. Resilient CPS are designed to withstand disruptions and remain\nfunctional despite the operation of adversaries. One of the dominant\nmethodologies explored for building resilient CPS is dependent on machine\nlearning (ML) algorithms. However, rising from recent research in adversarial\nML, we posit that ML algorithms for securing CPS must themselves be resilient.\nThis paper is therefore aimed at comprehensively surveying the interactions\nbetween resilient CPS using ML and resilient ML when applied in CPS. The paper\nconcludes with a number of research trends and promising future research\ndirections. Furthermore, with this paper, readers can have a thorough\nunderstanding of recent advances on ML-based security and securing ML for CPS\nand countermeasures, as well as research trends in this active research area.\n