English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Deep neural networks (DNNs) and, in particular, convolutional neural networks(CNNs) have brought significant advances in a wide range of modern computerapplication problems. However, the increasing availability of large amounts ofdatasets as well as the increasing available computational power of moderncomputers lead to a steady growth in the complexity and size of DNN and CNNmodels, and thus, to longer training times. Hence, various methods and attemptshave been developed to accelerate and parallelize the training of complexnetwork architectures. In this work, a novel CNN-DNN architecture is proposedthat naturally supports a model parallel training strategy and that is looselyinspired by two-level domain decomposition methods (DDM). First, local CNNmodels, that is, subnetworks, are defined that operate on overlapping ornonoverlapping parts of the input data, for example, sub-images. Thesubnetworks can be trained completely in parallel. Each subnetwork outputs alocal decision for the given machine learning problem which is exclusivelybased on the respective local input data. Subsequently, an additional DNN modelis trained which evaluates the local decisions of the local subnetworks andgenerates a final, global decision. With respect to the analogy to DDM, the DNNcan be interpreted as a coarse problem and hence, the new approach can beinterpreted as a two-level domain decomposition. In this paper, solely imageclassification problems using CNNs are considered. Experimental results fordifferent 2D image classification problems are provided as well as a facerecognition problem, and a classification problem for 3D computer tomography(CT) scans. The results show that the proposed approach can significantlyaccelerate the required training time compared to the global model and,additionally, can also help to improve the accuracy of the underlyingclassification problem.

A Domain Decomposition-Based CNN-DNN Architecture for Model Parallel Training Applied to Image Recognition Problems