Certifying Global Robustness for Deep Neural Networks

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Li, You, Zhao, Guannan, Kong, Shuyu, He, Yunqi, Zhou, Hai
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916266941349888
author Li, You
Zhao, Guannan
Kong, Shuyu
He, Yunqi
Zhou, Hai
author_facet Li, You
Zhao, Guannan
Kong, Shuyu
He, Yunqi
Zhou, Hai
contents A globally robust deep neural network resists perturbations on all meaningful inputs. Current robustness certification methods emphasize local robustness, struggling to scale and generalize. This paper presents a systematic and efficient method to evaluate and verify global robustness for deep neural networks, leveraging the PAC verification framework for solid guarantees on verification results. We utilize probabilistic programs to characterize meaningful input regions, setting a realistic standard for global robustness. Additionally, we introduce the cumulative robustness curve as a criterion in evaluating global robustness. We design a statistical method that combines multi-level splitting and regression analysis for the estimation, significantly reducing the execution time. Experimental results demonstrate the efficiency and effectiveness of our verification method and its capability to find rare and diversified counterexamples for adversarial training.
format Preprint
id arxiv_https___arxiv_org_abs_2405_20556
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Certifying Global Robustness for Deep Neural Networks
Li, You
Zhao, Guannan
Kong, Shuyu
He, Yunqi
Zhou, Hai
Machine Learning
Artificial Intelligence
A globally robust deep neural network resists perturbations on all meaningful inputs. Current robustness certification methods emphasize local robustness, struggling to scale and generalize. This paper presents a systematic and efficient method to evaluate and verify global robustness for deep neural networks, leveraging the PAC verification framework for solid guarantees on verification results. We utilize probabilistic programs to characterize meaningful input regions, setting a realistic standard for global robustness. Additionally, we introduce the cumulative robustness curve as a criterion in evaluating global robustness. We design a statistical method that combines multi-level splitting and regression analysis for the estimation, significantly reducing the execution time. Experimental results demonstrate the efficiency and effectiveness of our verification method and its capability to find rare and diversified counterexamples for adversarial training.
title Certifying Global Robustness for Deep Neural Networks
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2405.20556