DeepPath: Path-Driven Testing Criteria for Deep Neural Networks
2019 IEEE International Conference On Artificial Intelligence Testing (AITest)(2019)
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.
更多查看译文
关键词
deep neural networks, testing criteria, path coverage
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络