Multilayer feedforward networks are universal approximators
Neural Networks
POPL '77 Proceedings of the 4th ACM SIGACT-SIGPLAN symposium on Principles of programming languages
Counterexample-guided abstraction refinement for symbolic model checking
Journal of the ACM (JACM)
Developing artificial neural networks for safety critical systems
Neural Computing and Applications
Data Mining
Neural Networks: A Comprehensive Foundation (3rd Edition)
Neural Networks: A Comprehensive Foundation (3rd Edition)
Sketching concurrent data structures
Proceedings of the 2008 ACM SIGPLAN conference on Programming language design and implementation
The Journal of Machine Learning Research
Journal of Artificial Intelligence Research
Abstraction-guided synthesis of synchronization
Proceedings of the 37th annual ACM SIGPLAN-SIGACT symposium on Principles of programming languages
Neural networks for classification: a survey
IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews
NeVer: a tool for artificial neural networks verification
Annals of Mathematics and Artificial Intelligence
Challenging SMT solvers to verify neural networks
AI Communications - 18th RCRA International Workshop on “Experimental evaluation of algorithms for solving problems with combinatorial explosion”
| Hi-index | 0.00 |
A key problem in the adoption of artificial neural networks in safety-related applications is that misbehaviors can be hardly ruled out with traditional analytical or probabilistic techniques In this paper we focus on specific networks known as Multi-Layer Perceptrons (MLPs), and we propose a solution to verify their safety using abstractions to Boolean combinations of linear arithmetic constraints We show that our abstractions are consistent, i.e., whenever the abstract MLP is declared to be safe, the same holds for the concrete one Spurious counterexamples, on the other hand, trigger refinements and can be leveraged to automate the correction of misbehaviors We describe an implementation of our approach based on the HySAT solver, detailing the abstraction-refinement process and the automated correction strategy Finally, we present experimental results confirming the feasibility of our approach on a realistic case study.