Testing the Channels of Convolutional Neural Networks
This work addresses the challenge of ensuring correctness in CNNs for researchers and practitioners, but it is incremental as it builds on existing GAN methods for testing.
The paper tackles the problem of understanding and debugging convolutional neural networks (CNNs) by proposing techniques to test their channels, including FtGAN for generating test data and a channel selection algorithm, and shows that these techniques successfully identify defective channels in five CNN models across five public datasets.
Neural networks have complex structures, and thus it is hard to understand their inner workings and ensure correctness. To understand and debug convolutional neural networks (CNNs) we propose techniques for testing the channels of CNNs. We design FtGAN, an extension to GAN, that can generate test data with varying the intensity (i.e., sum of the neurons) of a channel of a target CNN. We also proposed a channel selection algorithm to find representative channels for testing. To efficiently inspect the target CNN's inference computations, we define unexpectedness score, which estimates how similar the inference computation of the test data is to that of the training data. We evaluated FtGAN with five public datasets and showed that our techniques successfully identify defective channels in five different CNN models.