Back to news

Learning-Based Approaches to Predictive Monitoring with Conformal Statistical Guarantees

Set 20, 2023

A tutorial paper in collaboration between King’s College London and University of Trieste on conformal prediction for predictive monitoring was accepted for publication at the 23rd International Conference on Runtime Verification that will take place in Thessaloniki (Greece) the next 3-6 October. This tutorial focuses on efficient methods to predictive monitoring (PM). The authors – Francesca Cairoli, Luca Bortolussi and Nicola Paoletti – review their recent work on predictive monitoring, one of the first to propose learning-based approximations for the verification of cyber-physical systems with respect to temporal logic specifications. In the tutorial, they present a general and comprehensive framework summarizing their predictive monitoring approach for several classes of systems, requirements, and assumptions.