CLUSTERING ALGORITHM NETWORKS TEST COST SENSITIVE FOR SPECIALIST DIVISIONS

ICTACT Journal on Image and Video Processing ( Volume: 10 , Issue: 2 )

Abstract

vioft2nntf2t|tblJournal|Abstract_paper|0xf4ff3a572b000000a2ad050001000200
It has been proven that deeper Convolutional Neural Networks (CNN) can result in better accuracy in many problems, but this accuracy comes with a high computational cost. Also, input instances have not the same difficulty. As a solution for accuracy vs. computational cost dilemma, we introduce a new test-cost-sensitive method for convolution neural networks. This method trains a CNN with a set Based on the difficulty of the input instance, the expert branches decide to use a shallower part of the network or go deeper to the end. The expert branches learn to determine: is the current network prediction wrong and if the instance passed to deeper network layers it will generate the right output; if not, then the expert branches will stop the process of computation. The experimental results on the standard CIFAR-10 dataset indicate that in comparison with basic models, the proposed method can train models with lower test cost and competitive accuracy.

Authors

M Arvindhan1, G S Pradeep Ghantasala2, N V Kousik3
Galgotias University, India1,3, Malla Reddy Institute of Technology and Science, India2

Keywords

Test-Cost-Sensitive Learning, Deep Learning, CNN with Expert Branches, Instance-based Cost

Published By
ICTACT
Published In
ICTACT Journal on Image and Video Processing
( Volume: 10 , Issue: 2 )
Date of Publication
November 2019
Pages
2098-2102

ICT Academy is an initiative of the Government of India in collaboration with the state Governments and Industries. ICT Academy is a not-for-profit society, the first of its kind pioneer venture under the Public-Private-Partnership (PPP) model

Contact Us

ICT Academy
Module No E6 -03, 6th floor Block - E
IIT Madras Research Park
Kanagam Road, Taramani,
Chennai 600 113,
Tamil Nadu, India

For Journal Subscription: journalsales@ictacademy.in

For further Queries and Assistance, write to us at: ictacademy.journal@ictacademy.in