Inducing Sparse Coding and And-Or Grammar from Generator Network

TL;DR

This paper presents a method to induce sparse coding and And-Or grammar in generator networks, enabling the learning of interpretable hierarchical image representations from primitive features to whole objects.

Contribution

It introduces a novel approach to enhance explainability in generative models by applying sparse operations and inducing hierarchical structures.

Findings

Learned meaningful hierarchical representations

Captured primitives, parts, and objects layer by layer

Produced explainable generative models

Abstract

We introduce an explainable generative model by applying sparse operation on the feature maps of the generator network. Meaningful hierarchical representations are obtained using the proposed generative model with sparse activations. The convolutional kernels from the bottom layer to the top layer of the generator network can learn primitives such as edges and colors, object parts, and whole objects layer by layer. From the perspective of the generator network, we propose a method for inducing both sparse coding and the AND-OR grammar for images. Experiments show that our method is capable of learning meaningful and explainable hierarchical representations.