Versatile Quantitative Modelling: Verification, Synthesis and Data Inference for Cyber-Physical Systems

摘要

Computing systems are becoming ever more complex, encompassing autonomous control of physical processes, stochasticity and inference from sensor data. This lecture will demonstrate the versatility of quantitative modelling and verification to aid the design of cyber-physical systems with machine learning components. Topics discussed will include recent advances in probabilistic/quantitative verification, template-based model synthesis, resource-performance trade off analysis, attacks on biometric security, and robustness guarantees for machine learning components. The lecture will conclude by giving an overview of future challenges in this field.