Improving Efficiency in Convolutional Neural Network with Multilinear Filters

TL;DR

This paper introduces a new neural network layer using multilinear projection that significantly reduces memory and computation needs while maintaining competitive performance compared to traditional CNNs.

Contribution

The paper proposes a novel neural network layer structure employing multilinear projection, offering a more efficient alternative to traditional CNN layers.

Findings

Requires several times less memory than CNNs

Outperforms traditional CNNs in experiments

Offers scalable computation schemes

Abstract

The excellent performance of deep neural networks has enabled us to solve several automatization problems, opening an era of autonomous devices. However, current deep net architectures are heavy with millions of parameters and require billions of floating point operations. Several works have been developed to compress a pre-trained deep network to reduce memory footprint and, possibly, computation. Instead of compressing a pre-trained network, in this work, we propose a generic neural network layer structure employing multilinear projection as the primary feature extractor. The proposed architecture requires several times less memory as compared to the traditional Convolutional Neural Networks (CNN), while inherits the similar design principles of a CNN. In addition, the proposed architecture is equipped with two computation schemes that enable computation reduction or scalability. Experimental results show the effectiveness of our compact projection that outperforms traditional CNN, while requiring far fewer parameters.