arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1902.05974cs.SEcs.CVcs.LG
收藏

DeepFault:深度神经网络的故障定位

DeepFault: Fault Localization for Deep Neural Networks

Bogazici University(博阿齐奇大学);University of York(约克大学)

Hasan Ferit Eniser, Simos Gerasimou, Alper Sen

摘要:

深度神经网络(DNN)越来越多地部署在安全关键型应用中,包括自动驾驶汽车和医疗诊断。为了降低意外DNN行为的剩余风险,并为其可靠运行提供证据,DNN应该进行彻底的测试。本文中提出的DeepFault白盒DNN测试方法通过采用受故障定位启发的可疑度量来建立神经元的命中谱并识别权重未正确校准的可疑神经元,从而解决了这一挑战,因此被认为是DNN性能不足的原因。DeepFault还使用了一种怀疑引导算法,从正确分类的输入中合成新的输入,增加可疑神经元的激活值。我们对在MNIST和CIFAR-10数据集上训练的几个DNN实例的经验评估表明,DeepFault在识别可疑神经元方面是有效的。此外,DeepFault合成的输入与原始输入非常相似,可以训练识别出的可疑神经元,并且具有高度的对抗性。

英文摘要:

Deep Neural Networks (DNNs) are increasingly deployed in safety-critical applications including autonomous vehicles and medical diagnostics. To reduce the residual risk for unexpected DNN behaviour and provide evidence for their trustworthy operation, DNNs should be thoroughly tested. The DeepFault whitebox DNN testing approach presented in our paper addresses this challenge by employing suspiciousness measures inspired by fault localization to establish the hit spectrum of neurons and identify suspicious neurons whose weights have not been calibrated correctly and thus are considered responsible for inadequate DNN performance. DeepFault also uses a suspiciousness-guided algorithm to synthesize new inputs, from correctly classified inputs, that increase the activation values of suspicious neurons. Our empirical evaluation on several DNN instances trained on MNIST and CIFAR-10 datasets shows that DeepFault is effective in identifying suspicious neurons. Also, the inputs synthesized by DeepFault closely resemble the original inputs, exercise the identified suspicious neurons and are highly adversarial.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑