A novel data-driven method for formal verification is proposed to study complex systems operating in safety-critical domains. The proposed approach is able to formally verify discrete-time stochastic dynamical systems against temporal logic specifications only using observation samples and without the knowledge of the model, and provide a probabilistic guarantee on the satisfaction of the specification. We first propose the theoretical results for using non-parametric estimation to estimate an asymptotic upper bound for the \emph{Lipschitz constant} of the stochastic system, which can determine a finite abstraction of the system. Our results prove that the asymptotic convergence rate of the estimation is $O(n^{-\frac{1}{3+d}})$, where $d$ is the dimension of the system and n is the data scale. We then construct interval Markov decision processes using two different data-driven methods, namely non-parametric estimation and empirical estimation of transition probabilities, to perform formal verification against a given temporal logic specification. Multiple case studies are presented to validate the effectiveness of the proposed methods.
History
Primary Research Area
Reliable Security Guarantees
Name of Conference
The 27th International Conference on Artificial Intelligence and Statistics
CISPA Affiliation
Yes
Journal
Proceedings of Machine Learning Research
Volume
238
Publisher
PMLR
BibTeX
@conference{Zhang:Ma:Soudijani:Soudjani:2024,
title = "Formal Verification of Unknown Stochastic Systems via Non-parametric Estimation",
author = "Zhang, Zhi" AND "Ma, Chenyu" AND "Soudijani, Saleh" AND "Soudjani, Sadegh",
year = 2024,
month = 5,
journal = "Proceedings of Machine Learning Research",
publisher = "PMLR"
}