Proj CDeepFuzz Paper Reading: DeepHunter: A Coverage-Guided Fuzz Testing Framework for Deep Neural Networks

Abstract

本文: DeepHunter
Task: Fuzzing Deep Learning Models
Github: https://github.com/Shimmer93/Deephunter-backup
Method:

1. Metamorphic mutation to generate new semantically preserved tests
2. use multiple plugable coverage criteria as feedback to guide the test generation
3. maintains multiple tests in a batch
4. prioritizes the test selection based on active feedback

实验：
datasets: MNIST(LeNet-1, LeNet-4, LeNet-5), CIFAR-10(ResNet-20, VGG-16), ImageNet(MobileNet, ResNet-50)
Feedback criteria:

1. Neuron Cov. (NC) The ratio of activated neurons;
2. K-multisec. Neu. Cov. (KMNC) The ratio of covered k-multisections of neurons;
3. Neuron Bound. Cov. (NBC) The ratio of covered boundary region of neurons;
4. Strong Neuron Act. Cov. (SNAC) The ratio of covered hyperactive boundary region;
5. Top-k Neu. Cov. (TKNC) The ratio of neurons in top-k hyperactived state;
6. Bottom-k Neu. Cov. (BKNC) The ratio of neurons in top-k hypoactived

效果：

1. 证明可以增加coverage
2. 能生成useful tests
3. 能为platform migration准确capture potential defects