Multiscale Hierarchical Convolutional Networks
This addresses the challenge of interpretability and efficiency in deep learning for image classification, though it appears incremental as it builds on existing convolutional network frameworks.
The paper tackles the problem of analyzing deep neural networks by introducing multiscale hierarchical convolutional networks, which achieve comparable precision to state-of-the-art networks on CIFAR image databases with significantly fewer parameters.
Deep neural network algorithms are difficult to analyze because they lack structure allowing to understand the properties of underlying transforms and invariants. Multiscale hierarchical convolutional networks are structured deep convolutional networks where layers are indexed by progressively higher dimensional attributes, which are learned from training data. Each new layer is computed with multidimensional convolutions along spatial and attribute variables. We introduce an efficient implementation of such networks where the dimensionality is progressively reduced by averaging intermediate layers along attribute indices. Hierarchical networks are tested on CIFAR image data bases where they obtain comparable precisions to state of the art networks, with much fewer parameters. We study some properties of the attributes learned from these databases.