Structural Compression of Convolutional Neural Networks

TL;DR

This paper presents CAR, a greedy filter pruning method for CNNs that produces smaller, more interpretable models with minimal accuracy loss by removing redundant filters and analyzing class importance.

Contribution

Introduces CAR, a novel filter pruning scheme that enhances CNN interpretability and reduces model size while maintaining accuracy.

Findings

CAR effectively prunes redundant filters like color filters.

Compressed CNNs retain diversity and accuracy.

Class importance analysis provides meaningful filter interpretations.

Abstract

Deep convolutional neural networks (CNNs) have been successful in many tasks in machine vision, however, millions of weights in the form of thousands of convolutional filters in CNNs makes them difficult for human intepretation or understanding in science. In this article, we introduce CAR, a greedy structural compression scheme to obtain smaller and more interpretable CNNs, while achieving close to original accuracy. The compression is based on pruning filters with the least contribution to the classification accuracy. We demonstrate the interpretability of CAR-compressed CNNs by showing that our algorithm prunes filters with visually redundant functionalities such as color filters. These compressed networks are easier to interpret because they retain the filter diversity of uncompressed networks with order of magnitude less filters. Finally, a variant of CAR is introduced to quantify the importance of each image category to each CNN filter. Specifically, the most and the least important class labels are shown to be meaningful interpretations of each filter.