Speeding up neural network robustness verification via algorithm configuration and an optimised mixed integer linear programming solver portfolio

Machine Learning, vol. 111, pp. 4565–4584

Abstract

Abstract Despite their great success in recent years, neural networks have been found to be vulnerable to adversarial attacks. These attacks are often based on slight perturbations of given inputs that cause them to be misclassified. Several methods have been proposed to formally prove robustness of a given network against such attacks. However, these methods typically give rise to high computational demands, which severely limit their scalability. Recent state-of-the-art approaches state the verification task as a minimisation problem, which is formulated and solved as a mixed-integer linear programming (MIP) problem. We extend this approach by leveraging automated algorithm configuration techniques and, more specifically, construct a portfolio of MIP solver configurations optimised for the neural network verification task. We test this approach on two recent, state-of-the-art MIP-based verification engines, $$\mathrm {MIPVerify}$$ MIPVerify and $$\mathrm {Venus}$$ Venus , and achieve substantial improvements in CPU time by average factors of up to 4.7 and 10.3, respectively.

Authors 2

1. Holger H. Hoos Aachen Leiden Institute of Advanced Computer Science

   Leiden University · University of British Columbia

   Affiliation as printed

   Leiden Institute of Advanced Computer Science, Leiden University, Leiden, The Netherlands

   University of British Columbia, Vancouver, Canada

2. Jan N. van Rijn Aachen Leiden Institute of Advanced Computer Science

   Leiden University

   Affiliation as printed

   Leiden Institute of Advanced Computer Science, Leiden University, Leiden, The Netherlands