Property-Directed Verification of Recurrent Neural Networks - Archive ouverte HAL
Preprints, Working Papers, ... Year : 2020

Property-Directed Verification of Recurrent Neural Networks

Abstract

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.

Dates and versions

hal-03134999 , version 1 (08-02-2021)

Identifiers

Cite

Igor Khmelnitsky, Daniel Neider, Rajarshi Roy, Benoît Barbot, Benedikt Bollig, et al.. Property-Directed Verification of Recurrent Neural Networks. 2020. ⟨hal-03134999⟩
148 View
0 Download

Altmetric

Share

More