DeepPath: Path-driven Testing Criteria for Deep Neural Networks

被引：12

作者：

Wang, Dong ^{[1
]}

Wang, Ziyuan ^{[2
]}

Fang, Chunrong ^{[1
]}

Chen, Yanshan ^{[2
]}

Chen, Zhenyu ^{[1
]}

机构：

[1] Nanjing Univ, Inst Software, Nanjing, Jiangsu, Peoples R China

[2] Nanjing Univ Posts & Telecommun, Sch Comp Sci, Nanjing, Jiangsu, Peoples R China

来源：

2019 IEEE INTERNATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE TESTING (AITEST) | 2019年

基金：

中国国家自然科学基金;

关键词：

deep neural networks; testing criteria; path coverage;

D O I：

10.1109/AITest.2019.00013

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Inspired by path-oriented testing, we propose a series of path-driven testing criteria, called DeepPath, to comprehensively calculate coverage in deep neural networks (DNNs). Four DNN models and four adversarial attack techniques are used to evaluate the effectiveness of DeepPath. The experimental results illustrate that DeepPath are more discriminating to measure test adequacy of DNNs in practice, as well as more useful for recognizing adversarial attack test inputs.

引用

页码：119 / 120

页数：2