GA-based filter selection for representation in convolutional neural networks

Junbong Kim; Minki Lee; Jongeun Choi; Kisung Seo

doi:10.1007/978-3-030-11018-5_48

GA-based filter selection for representation in convolutional neural networks

Junbong Kim, Minki Lee, Jongeun Choi, Kisung Seo

Department of Mechanical Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

One of the deep learning models, a convolutional neural network (CNN) has been very successful in a variety of computer vision tasks. Features of a CNN are automatically generated, however, they can be further optimized since they often require large scale parallel operations and there exist the possibility of overlapping redundant features. The aim of this paper is to use feature selection via evolutionary algorithms to remove the irrelevant deep features. This will minimize the computational complexity and the amount of overfitting while maintaining a good quality of representation. We demonstrate the improvement of the filter representation by performing experiments on three data sets of CIFAR10, metal surface defects, and variation of MNIST and by analyzing the classification performance as well as the variance of the filter.

Original language	English
Title of host publication	Computer Vision – ECCV 2018 Workshops, Proceedings
Editors	Laura Leal-Taixé, Stefan Roth
Publisher	Springer Verlag
Pages	609-618
Number of pages	10
ISBN (Print)	9783030110178
DOIs	https://doi.org/10.1007/978-3-030-11018-5_48
Publication status	Published - 2019
Event	15th European Conference on Computer Vision, ECCV 2018 - Munich, Germany Duration: 2018 Sep 8 → 2018 Sep 14

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11132 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Other

Other	15th European Conference on Computer Vision, ECCV 2018
Country	Germany
City	Munich
Period	18/9/8 → 18/9/14

Bibliographical note

Funding Information:
Acknowledgement. This work was supported by National Research Foundation of Korea Grant funded by the Korea government (NRF-2016R1D1A1A09919650). This work was also funded by the Korea Meteorological Administration Research and Development Program under Grant KMIPA(KMI2018-06710).

Publisher Copyright:
© Springer Nature Switzerland AG 2019.

All Science Journal Classification (ASJC) codes

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/978-3-030-11018-5_48

Fingerprint Dive into the research topics of 'GA-based filter selection for representation in convolutional neural networks'. Together they form a unique fingerprint.

Cite this

Kim, J., Lee, M., Choi, J., & Seo, K. (2019). GA-based filter selection for representation in convolutional neural networks. In L. Leal-Taixé, & S. Roth (Eds.), Computer Vision – ECCV 2018 Workshops, Proceedings (pp. 609-618). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11132 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-11018-5_48

Kim, Junbong ; Lee, Minki ; Choi, Jongeun ; Seo, Kisung. / GA-based filter selection for representation in convolutional neural networks. Computer Vision – ECCV 2018 Workshops, Proceedings. editor / Laura Leal-Taixé ; Stefan Roth. Springer Verlag, 2019. pp. 609-618 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{bfa7c10dcaaf4684b2e8ca161758e52d,

title = "GA-based filter selection for representation in convolutional neural networks",

abstract = "One of the deep learning models, a convolutional neural network (CNN) has been very successful in a variety of computer vision tasks. Features of a CNN are automatically generated, however, they can be further optimized since they often require large scale parallel operations and there exist the possibility of overlapping redundant features. The aim of this paper is to use feature selection via evolutionary algorithms to remove the irrelevant deep features. This will minimize the computational complexity and the amount of overfitting while maintaining a good quality of representation. We demonstrate the improvement of the filter representation by performing experiments on three data sets of CIFAR10, metal surface defects, and variation of MNIST and by analyzing the classification performance as well as the variance of the filter.",

author = "Junbong Kim and Minki Lee and Jongeun Choi and Kisung Seo",

note = "Funding Information: Acknowledgement. This work was supported by National Research Foundation of Korea Grant funded by the Korea government (NRF-2016R1D1A1A09919650). This work was also funded by the Korea Meteorological Administration Research and Development Program under Grant KMIPA(KMI2018-06710). Publisher Copyright: {\textcopyright} Springer Nature Switzerland AG 2019.; 15th European Conference on Computer Vision, ECCV 2018 ; Conference date: 08-09-2018 Through 14-09-2018",

year = "2019",

doi = "10.1007/978-3-030-11018-5_48",

language = "English",

isbn = "9783030110178",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "609--618",

editor = "Laura Leal-Taix{\'e} and Stefan Roth",

booktitle = "Computer Vision – ECCV 2018 Workshops, Proceedings",

address = "Germany",

}

Kim, J, Lee, M, Choi, J & Seo, K 2019, GA-based filter selection for representation in convolutional neural networks. in L Leal-Taixé & S Roth (eds), Computer Vision – ECCV 2018 Workshops, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11132 LNCS, Springer Verlag, pp. 609-618, 15th European Conference on Computer Vision, ECCV 2018, Munich, Germany, 18/9/8. https://doi.org/10.1007/978-3-030-11018-5_48

GA-based filter selection for representation in convolutional neural networks. / Kim, Junbong; Lee, Minki; Choi, Jongeun; Seo, Kisung.

Computer Vision – ECCV 2018 Workshops, Proceedings. ed. / Laura Leal-Taixé; Stefan Roth. Springer Verlag, 2019. p. 609-618 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11132 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - GA-based filter selection for representation in convolutional neural networks

AU - Kim, Junbong

AU - Lee, Minki

AU - Choi, Jongeun

AU - Seo, Kisung

N1 - Funding Information: Acknowledgement. This work was supported by National Research Foundation of Korea Grant funded by the Korea government (NRF-2016R1D1A1A09919650). This work was also funded by the Korea Meteorological Administration Research and Development Program under Grant KMIPA(KMI2018-06710). Publisher Copyright: © Springer Nature Switzerland AG 2019.

PY - 2019

Y1 - 2019

N2 - One of the deep learning models, a convolutional neural network (CNN) has been very successful in a variety of computer vision tasks. Features of a CNN are automatically generated, however, they can be further optimized since they often require large scale parallel operations and there exist the possibility of overlapping redundant features. The aim of this paper is to use feature selection via evolutionary algorithms to remove the irrelevant deep features. This will minimize the computational complexity and the amount of overfitting while maintaining a good quality of representation. We demonstrate the improvement of the filter representation by performing experiments on three data sets of CIFAR10, metal surface defects, and variation of MNIST and by analyzing the classification performance as well as the variance of the filter.

AB - One of the deep learning models, a convolutional neural network (CNN) has been very successful in a variety of computer vision tasks. Features of a CNN are automatically generated, however, they can be further optimized since they often require large scale parallel operations and there exist the possibility of overlapping redundant features. The aim of this paper is to use feature selection via evolutionary algorithms to remove the irrelevant deep features. This will minimize the computational complexity and the amount of overfitting while maintaining a good quality of representation. We demonstrate the improvement of the filter representation by performing experiments on three data sets of CIFAR10, metal surface defects, and variation of MNIST and by analyzing the classification performance as well as the variance of the filter.

UR - http://www.scopus.com/inward/record.url?scp=85061737419&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85061737419&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-11018-5_48

DO - 10.1007/978-3-030-11018-5_48

M3 - Conference contribution

AN - SCOPUS:85061737419

SN - 9783030110178

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 609

EP - 618

BT - Computer Vision – ECCV 2018 Workshops, Proceedings

A2 - Leal-Taixé, Laura

A2 - Roth, Stefan

PB - Springer Verlag

T2 - 15th European Conference on Computer Vision, ECCV 2018

Y2 - 8 September 2018 through 14 September 2018

ER -

Kim J, Lee M, Choi J, Seo K. GA-based filter selection for representation in convolutional neural networks. In Leal-Taixé L, Roth S, editors, Computer Vision – ECCV 2018 Workshops, Proceedings. Springer Verlag. 2019. p. 609-618. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-11018-5_48

GA-based filter selection for representation in convolutional neural networks

Abstract

Publication series

Other

Bibliographical note

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Cite this