2020/09/22 by Igor Khmelnitsky, Khmelnitsky, Igor, Daniel Neider +15 · 1 citation
Computer Science · #68Q60 #68T07 #Adversarial Robustness in Machine Learning #Artificial Intelligence (cs.AI) #D.2.4 #FOS: Computer and information sciences #Formal Languages and Automata Theory (cs.FL) #Formal Methods in Verification #I.2.6 #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Machine Learning and Algorithms
paper · pdf · doi:10.48550/arxiv.2009.10610
openalex publication_date 2020/09/22 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
This paper presents a property-directed approach to verifying recurrent neural networks (RNNs). To this end, we learn a deterministic finite automaton as a surrogate model from a given RNN using active automata learning. This model may then be analyzed using model checking as verification technique. The term property-directed reflects the idea that our procedure is guided and controlled by the given property rather than performing the two steps separately. We show that this not only allows us to discover small counterexamples fast, but also to generalize them by pumping towards faulty flows hinting at the underlying error in the RNN.