Saved in:
Bibliographic Details
Main Authors: Singhal, Utkarsh, Xing, Yifei, Yu, Stella X.
Format: Preprint
Published: 2021
Subjects:
Online Access:https://arxiv.org/abs/2112.01525
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915253518860288
author Singhal, Utkarsh
Xing, Yifei
Yu, Stella X.
author_facet Singhal, Utkarsh
Xing, Yifei
Yu, Stella X.
contents We study complex-valued scaling as a type of symmetry natural and unique to complex-valued measurements and representations. Deep Complex Networks (DCN) extends real-valued algebra to the complex domain without addressing complex-valued scaling. SurReal takes a restrictive manifold view of complex numbers, adopting a distance metric to achieve complex-scaling invariance while losing rich complex-valued information. We analyze complex-valued scaling as a co-domain transformation and design novel equivariant and invariant neural network layer functions for this special transformation. We also propose novel complex-valued representations of RGB images, where complex-valued scaling indicates hue shift or correlated changes across color channels. Benchmarked on MSTAR, CIFAR10, CIFAR100, and SVHN, our co-domain symmetric (CDS) classifiers deliver higher accuracy, better generalization, robustness to co-domain transformations, and lower model bias and variance than DCN and SurReal with far fewer parameters.
format Preprint
id arxiv_https___arxiv_org_abs_2112_01525
institution arXiv
publishDate 2021
record_format arxiv
spellingShingle Co-domain Symmetry for Complex-Valued Deep Learning
Singhal, Utkarsh
Xing, Yifei
Yu, Stella X.
Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
We study complex-valued scaling as a type of symmetry natural and unique to complex-valued measurements and representations. Deep Complex Networks (DCN) extends real-valued algebra to the complex domain without addressing complex-valued scaling. SurReal takes a restrictive manifold view of complex numbers, adopting a distance metric to achieve complex-scaling invariance while losing rich complex-valued information. We analyze complex-valued scaling as a co-domain transformation and design novel equivariant and invariant neural network layer functions for this special transformation. We also propose novel complex-valued representations of RGB images, where complex-valued scaling indicates hue shift or correlated changes across color channels. Benchmarked on MSTAR, CIFAR10, CIFAR100, and SVHN, our co-domain symmetric (CDS) classifiers deliver higher accuracy, better generalization, robustness to co-domain transformations, and lower model bias and variance than DCN and SurReal with far fewer parameters.
title Co-domain Symmetry for Complex-Valued Deep Learning
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2112.01525