Hybrid Quantum-Classical Neural Network for Cloud-supported In-Vehicle Cyberattack Detection

A classical computer works with ones and zeros, whereas a quantum uses ones, superpositions of which enables computers to perform vast number calculations simultaneously compared computers. In cloud-supported cyber−physical system environment, running machine learning application in is often difficult, due the existing limitations current devices. However, combination quantum-classical neural networks (NN), complex high-dimensional features can be extracted by NN reduced but more informative feature space processed this study, we developed hybrid detect an amplitude shift cyberattack on in-vehicle controller area network dataset. We showed that using NN, it possible achieve attack detection accuracy 94%, higher than long short-term memory (88%) or alone (62%).