2019/07/05 by Michel Barbeau, Barbeau, Michel, Joaquín García-Alfaro +1
Computer Science · #Advanced Malware Detection Techniques #Adversarial Robustness in Machine Learning #Chaos-based Image/Signal Encryption #Cryptography and Security (cs.CR) #FOS: Computer and information sciences #Machine Learning (cs.LG)
paper · pdf · doi:10.48550/arxiv.1907.03038
openalex publication_date 2019/07/05 · openalex created_date 2022/07/28 · openalex updated_date 2026/07/28
We show that the Quantum Generative Adversarial Network (QGAN) paradigm can\nbe employed by an adversary to learn generating data that deceives the\nmonitoring of a Cyber-Physical System (CPS) and to perpetrate a covert attack.\nAs a test case, the ideas are elaborated considering the navigation data of a\nMicro Aerial Vehicle (MAV). A concrete QGAN design is proposed to generate fake\nMAV navigation data. Initially, the adversary is entirely ignorant about the\ndynamics of the CPS, the strength of the approach from the point of view of the\nbad guy. A design is also proposed to discriminate between genuine and fake MAV\nnavigation data. The designs combine classical optimization, qubit quantum\ncomputing and photonic quantum computing. Using the PennyLane software\nsimulation, they are evaluated over a classical computing platform. We assess\nthe learning time and accuracy of the navigation data generator and\ndiscriminator versus space complexity, i.e., the amount of quantum memory\nneeded to solve the problem.\n