Safety Verification of Tree-Ensemble Policies via Predicate Abstraction - LAAS - Laboratoire d'Analyse et d'Architecture des Systèmes
Communication Dans Un Congrès Année : 2024

Safety Verification of Tree-Ensemble Policies via Predicate Abstraction

Résumé

Learned action policies are gaining traction in AI, but come without safety guarantees. Recent work devised a method for safety verification of neural policies via predicate abstraction. Here we extend this approach to policies represented by tree ensembles, through replacing the underlying SMT queries with queries that can be dispatched by Veritas, a reasoning tool dedicated to tree ensembles. The query language supported by Veritas is limited, and we show how to encode richer constraints we need into additional trees and decision variables. We run experiments on benchmarks previously used to evaluate neural policy verification, and we design new benchmarks based on a logistics application at Airbus as well as on a real-world robotics domain. We find that (1) verification with Veritas vastly outperforms verification with Z3 and Gurobi; (2) treeensemble policies are much faster to verify than neural policies, while being competitive in policy quality; (3) our techniques are highly complementary to, and often outperform, an encoding of treeensemble policy verification into NUXMV.
Fichier principal
Vignette du fichier
ecai24.pdf (324.38 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04822331 , version 1 (06-12-2024)

Identifiants

Citer

Chaahat Jain, Lorenzo Cascioli, Laurens Devos, Marcel Vinzent, Marcel Steinmetz, et al.. Safety Verification of Tree-Ensemble Policies via Predicate Abstraction. ECAI 2024 - 27th European Conference on Artificial Intelligence, Oct 2024, Santiago de Compostela, Spain. ⟨10.3233/FAIA240614⟩. ⟨hal-04822331⟩
0 Consultations
0 Téléchargements

Altmetric

Partager

More