Theoretical Properties for Neural Networks with Weight Matrices of Low Displacement Rank
Liang Zhao, Siyu Liao, Yanzhi Wang, Zhe Li, Jian Tang, Bo Yuan
Abstract
Recently low displacement rank (LDR) matrices, or so-called structured matrices, have been proposed to compress large-scale neural networks. Empirical results have shown that neural networks with weight matrices of LDR matrices, referred as LDR neural networks, can achieve significant reduction in space and computational complexity while retaining high accuracy. This paper gives theoretical study on LDR neural networks. First, we prove the universal approximation property of LDR neural networks with a mild condition on the displacement operators. We then show that the error bounds of LDR neural networks are as efficient as general neural networks with both single-layer and multiple-layer structure. Finally, we propose back-propagation based training algorithm for general LDR neural networks.
BibTeX
@InProceedings{pmlr-v70-zhao17b,
title = {Theoretical Properties for Neural Networks with Weight Matrices of Low Displacement Rank},
author = {Liang Zhao and Siyu Liao and Yanzhi Wang and Zhe Li and Jian Tang and Bo Yuan},
booktitle = {Proceedings of the 34th International Conference on Machine Learning},
pages = {4082--4090},
year = {2017},
editor = {Precup, Doina and Teh, Yee Whye},
volume = {70},
series = {Proceedings of Machine Learning Research},
month = {06--11 Aug},
publisher = {PMLR},
pdf = {http://proceedings.mlr.press/v70/zhao17b/zhao17b.pdf},
url = {https://proceedings.mlr.press/v70/zhao17b.html},
abstract = {Recently low displacement rank (LDR) matrices, or so-called structured matrices, have been proposed to compress large-scale neural networks. Empirical results have shown that neural networks with weight matrices of LDR matrices, referred as LDR neural networks, can achieve significant reduction in space and computational complexity while retaining high accuracy. This paper gives theoretical study on LDR neural networks. First, we prove the universal approximation property of LDR neural networks with a mild condition on the displacement operators. We then show that the error bounds of LDR neural networks are as efficient as general neural networks with both single-layer and multiple-layer structure. Finally, we propose back-propagation based training algorithm for general LDR neural networks.}
}