Towards General Robustness Verification of MaxPool-based Convolutional Neural Networks via Tightening Linear Approximation

TL;DR

This paper introduces MaxLin, a verification tool that tightens linear approximations of MaxPool functions in CNNs, significantly improving the certified robustness bounds and efficiency over existing methods.

Contribution

It presents a novel approach to tighten linear approximations of MaxPool, enabling more accurate and scalable robustness verification of MaxPool-based CNNs.

Findings

MaxLin achieves up to 110.60% improvement in certified lower bounds.

MaxLin provides up to 5.13× speedup compared to state-of-the-art tools.

Evaluation on benchmarks shows superior performance across datasets.

Abstract

The robustness of convolutional neural networks (CNNs) is vital to modern AI-driven systems. It can be quantified by formal verification by providing a certified lower bound, within which any perturbation does not alter the original input's classification result. It is challenging due to nonlinear components, such as MaxPool. At present, many verification methods are sound but risk losing some precision to enhance efficiency and scalability, and thus, a certified lower bound is a crucial criterion for evaluating the performance of verification tools. In this paper, we present MaxLin, a robustness verifier for MaxPool-based CNNs with tight linear approximation. By tightening the linear approximation of the MaxPool function, we can certify larger certified lower bounds of CNNs. We evaluate MaxLin with open-sourced benchmarks, including LeNet and networks trained on the MNIST, CIFAR-10, and Tiny ImageNet datasets. The results show that MaxLin outperforms state-of-the-art tools with up to 110.60% improvement regarding the certified lower bound and 5.13 $\times$ speedup for the same neural networks. Our code is available at https://github.com/xiaoyuanpigo/maxlin.