Proj CDeepFuzz Paper Reading: Testing Deep Neural Networks

Abstract

本文:DeepCover
Github: https://github.com/TrustAI/DeepCover
Task: propose 4 novel test criteria to test DNNs
Method: inspired by MC/DC coverage criteria

1. Sign-Sign Coverage(Main): the sign change of a condition feature ψk,independently affects the sign of the decision feature ψk+1,j of the next layer
2. Value-Sign Coverage:
3. Sign-Value Coverage: captures the significant change of a decision feature’s value that complements the sign change case
4. Value-Value Coverage: there is no sign change for a condition feature, but the decision feature’s value is changed significantly

SubTask: adversarial testing: symbolic method, gradient-based heuristic search

实验：
数据集：MNIST, CIFAR-10, ImageNet
效果：

1. 通过这些criteria得到的test samples能够诱发DNN的undesired behavior
2. 在finding bugs与计算代价之间达到了平衡

本栏目推荐文章
• 10.基于模型的测试方法 Model-based Testing
• Generalised f-Mean Aggregation for Graph Neural Networks
• Cannot read properties of null (reading 'parentElement')问题的解决
• Graph Condensation for Graph Neural Networks
• 论文精读：ST2Vec：道路网络中的时空轨迹相似性学习（ST2Vec: Spatio_Temporal Trajectory Similarity Learning in Road Networks）
• 论文阅读-Self-supervised and Interpretable Data Cleaning with Sequence Generative Adversarial Networks
• 如何在谷歌 Google Chrome 浏览器禁用掉右键菜单中的【使用朗读模式 Reading Mode 打开】
• BIgdataAIML-IBM-A neural networks deep dive - An introduction to neural networks and their programming
• Relation Networks for Object Detection
• Local Relation Networks for Image Recognition： LRNet

CDeepFuzz Networks Reading Testing Neuralcdeepfuzz networks reading testing deep-neural-network-driven autonomous cdeepfuzz learning networks neural hello combinatorial cdeepfuzz learning reading differentiation cdeepfuzz automatic reading pre-trained cdeepfuzz natural reading graph condensation networks neural understanding plasticity networks neural networks neural bigdataaiml-ibm-a introduction decoupling networks neural depth