Computer Science > Numerical Analysis
[Submitted on 20 May 2018 (this version), latest version 3 Oct 2018 (v3)]
Title:Low-Cost Parameterizations of Deep Convolution Neural Networks
View PDFAbstract:The main computational cost in the training of and prediction with Convolution Neural Networks (CNNs) typically stems from the convolution. In this paper, we present three novel ways to parameterize the convolution more efficiently, significantly decreasing the computational complexity. Commonly used CNNs filter the input data using a series of spatial convolutions with compact stencils that couple features from all channels and point-wise nonlinearities. In this paper, we propose three architectures that are cheaper to couple the channel dimension and thereby reduce both the number of trainable weights and the computational cost of the CNN. The first architecture is inspired by tensor-products and imposes a circulant coupling of the channels. The second and third architectures arise as discretizations of a new type of residual neural network (ResNet) that is inspired by Partial Differential Equations (PDEs) or reaction-diffusion type. The coupling patterns of the first two architectures are applicable to a large class of CNNs. We outline in our numerical experiments that the proposed architectures, although considerably reducing the number of trainable weights, yield comparable accuracy to existing CNNs that are fully coupled in the channel dimension.
Submission history
From: Lars Ruthotto [view email][v1] Sun, 20 May 2018 20:17:47 UTC (47 KB)
[v2] Tue, 2 Oct 2018 07:32:58 UTC (50 KB)
[v3] Wed, 3 Oct 2018 06:10:32 UTC (23 KB)
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.