Lossless Compression of Deep Neural Networks

TL;DR

This paper presents LEO, a lossless neural network compression algorithm that uses MILP and L1 regularization to remove units and layers without affecting the network's output, aiding deployment on resource-limited devices.

Contribution

Introduction of LEO, a novel lossless compression method for neural networks that leverages MILP and L1 regularization to optimize network size without accuracy loss.

Findings

LEO can compress neural networks without changing their outputs.

The method effectively identifies linear ReLUs using MILP.

L1 regularization facilitates larger initial architectures for better compression.

Abstract

Deep neural networks have been successful in many predictive modeling tasks, such as image and language recognition, where large neural networks are often used to obtain good accuracy. Consequently, it is challenging to deploy these networks under limited computational resources, such as in mobile devices. In this work, we introduce an algorithm that removes units and layers of a neural network while not changing the output that is produced, which thus implies a lossless compression. This algorithm, which we denote as LEO (Lossless Expressiveness Optimization), relies on Mixed-Integer Linear Programming (MILP) to identify Rectified Linear Units (ReLUs) with linear behavior over the input domain. By using L1 regularization to induce such behavior, we can benefit from training over a larger architecture than we would later use in the environment where the trained neural network is deployed.