Visualization, Discriminability and Applications of Interpretable Saak Features
This work addresses the problem of interpretability in deep learning for researchers and practitioners, but it is incremental as it builds on existing Saak transform methods.
The paper tackled the challenge of interpretable deep learning by analyzing Saak features, demonstrating their discriminative power and achieving competitive classification performance on MNIST, CIFAR-10, and STL-10 datasets.
In this work, we study the power of Saak features as an effort towards interpretable deep learning. Being inspired by the operations of convolutional layers of convolutional neural networks, multi-stage Saak transform was proposed. Based on this foundation, we provide an in-depth examination on Saak features, which are coefficients of the Saak transform, by analyzing their properties through visualization and demonstrating their applications in image classification. Being similar to CNN features, Saak features at later stages have larger receptive fields, yet they are obtained in a one-pass feedforward manner without backpropagation. The whole feature extraction process is transparent and is of extremely low complexity. The discriminant power of Saak features is demonstrated, and their classification performance in three well-known datasets (namely, MNIST, CIFAR-10 and STL-10) is shown by experimental results.