Saved in:
Bibliographic Details
Main Authors: Kiruluta, Andrew, Williams, Samantha
Format: Preprint
Published: 2017
Subjects:
Online Access:https://arxiv.org/abs/1801.01451
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916694878846976
author Kiruluta, Andrew
Williams, Samantha
author_facet Kiruluta, Andrew
Williams, Samantha
contents This paper presents a Sparse Hierarchical Fourier Interaction Networks, an architectural building block that unifies three complementary principles of frequency domain modeling: A hierarchical patch wise Fourier transform that affords simultaneous access to local detail and global context; A learnable, differentiable top K masking mechanism which retains only the most informative spectral coefficients, thereby exploiting the natural compressibility of visual and linguistic signals.
format Preprint
id arxiv_https___arxiv_org_abs_1801_01451
institution arXiv
publishDate 2017
record_format arxiv
spellingShingle Reducing Deep Network Complexity via Sparse Hierarchical Fourier Interaction Networks
Kiruluta, Andrew
Williams, Samantha
Computer Vision and Pattern Recognition
Machine Learning
This paper presents a Sparse Hierarchical Fourier Interaction Networks, an architectural building block that unifies three complementary principles of frequency domain modeling: A hierarchical patch wise Fourier transform that affords simultaneous access to local detail and global context; A learnable, differentiable top K masking mechanism which retains only the most informative spectral coefficients, thereby exploiting the natural compressibility of visual and linguistic signals.
title Reducing Deep Network Complexity via Sparse Hierarchical Fourier Interaction Networks
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/1801.01451