Tampere University of Technology

TUTCRIS Research Portal

Improving efficiency in convolutional neural networks with multilinear filters

Research output: Contribution to journalArticleScientificpeer-review

Details

Original languageEnglish
Pages (from-to)328-339
Number of pages12
JournalNeural Networks
Volume105
DOIs
Publication statusPublished - 1 Sep 2018
Publication typeA1 Journal article-refereed

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.

Keywords

  • Convolutional neural networks, Multilinear projection, Network compression

Publication forum classification

Field of science, Statistics Finland