NoiseOut: A Simple Way to Prune Neural Networks

TL;DR

NoiseOut is an automated neural network pruning method that leverages neuron activation correlations, enhanced by adding random output neurons, achieving high pruning rates without sacrificing accuracy.

Contribution

We introduce NoiseOut, a novel pruning algorithm that uses activation correlation and random output neurons to improve pruning efficiency and effectiveness.

Findings

High pruning rates achieved across various networks

Maintains original network accuracy after pruning

Adding random output neurons increases pruning effectiveness

Abstract

Neural networks are usually over-parameterized with significant redundancy in the number of required neurons which results in unnecessary computation and memory usage at inference time. One common approach to address this issue is to prune these big networks by removing extra neurons and parameters while maintaining the accuracy. In this paper, we propose NoiseOut, a fully automated pruning algorithm based on the correlation between activations of neurons in the hidden layers. We prove that adding additional output neurons with entirely random targets results into a higher correlation between neurons which makes pruning by NoiseOut even more efficient. Finally, we test our method on various networks and datasets. These experiments exhibit high pruning rates while maintaining the accuracy of the original network.