Critically Assessing the State of the Art in Neural Network Verification

Recent research has proposed various methods to formally verify neural networks against minimal input perturbations; this verification task is also known as local robustness verification. The research area of local robustness verification is highly diverse, as verifiers rely on a multitude of techniques, including mixed integer programming and satisfiability modulo theories. At the same time, the problem instances encountered when performing local robustness verification differ based on the network to be verified, the property to be verified and the specific …

art diverse mixed network networks neural network neural networks programming research robustness state state of the art verification verify