Resilient Machine Learning for Networked Cyber Physical Systems: A Survey for Machine Learning Security to Securing Machine Learning for CPS

Cyber Physical Systems (CPS) are characterized by their ability to integrate the physical and information or cyber worlds. Their deployment in critical infrastructure have demonstrated a potential transform world. However, harnessing this is limited nature far reaching effects of attacks on human, environment. An attraction for concerns CPS rises from process sending sensors actuators over wireless communication medium, thereby widening attack surface. Traditionally, security has been investigated perspective preventing intruders gaining access system using cryptography other control techniques. Most research work therefore focused detection CPS. world increasing adversaries, it becoming more difficult totally prevent adversarial attacks, hence need focus making resilient. Resilient designed withstand disruptions remain functional despite operation adversaries. One dominant methodologies explored building resilient dependent machine learning (ML) algorithms. rising recent ML, we posit that ML algorithms securing must themselves be This article aimed at comprehensively surveying interactions between when applied The paper concludes with number trends promising future directions. Furthermore, article, readers can thorough understanding advances ML-based countermeasures, as well active area.