Neural Architecture Search for Extreme Multi-label Text Classification

Loïc Pauletto; Massih Reza Amini; Rohit Babbar; Nicolas Winckler

doi:10.1007/978-3-030-63836-8_24

Neural Architecture Search for Extreme Multi-label Text Classification

Loïc Pauletto, Massih Reza Amini, Rohit Babbar, Nicolas Winckler

Department of Computer Science

Research output: Chapter or section in a book/report/conference proceeding › Chapter in a published conference proceeding

2 Citations (SciVal)

Abstract

Extreme classification and Neural Architecture Search (NAS) are research topics which have recently gained a lot of interest. While the former has been mainly motivated and applied in e-commerce and Natural Language Processing (NLP) applications, the NAS approach has been applied to a small variety of tasks, mainly in image processing. In this study, we extend the scope of NAS to the task of extreme multilabel classification (XMC). We propose a neuro-evolution approach, which was found to be the most suitable for a variety of tasks. Our NAS method automatically finds architectures that give competitive results with respect to the state of the art (and superior to other methods) with faster convergence. In addition, we perform analysis of the weights of the architecture blocks to provide insight into the importance of different operations that have been selected by the method.

Original language	English
Title of host publication	Neural Information Processing - 27th International Conference, ICONIP 2020, Proceedings
Editors	Haiqin Yang, Kitsuchart Pasupa, Andrew Chi-Sing Leung, James T. Kwok, Jonathan H. Chan, Irwin King
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	282-293
Number of pages	12
ISBN (Print)	9783030638351
DOIs	https://doi.org/10.1007/978-3-030-63836-8_24
Publication status	Published - 19 Nov 2020
Event	27th International Conference on Neural Information Processing, ICONIP 2020 - Bangkok, Thailand Duration: 18 Nov 2020 → 22 Nov 2020

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12534 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	27th International Conference on Neural Information Processing, ICONIP 2020
Country/Territory	Thailand
City	Bangkok
Period	18/11/20 → 22/11/20

Bibliographical note

Publisher Copyright:
© 2020, Springer Nature Switzerland AG.

Keywords

Evolutionary algorithms
Extreme multi-label text classification
Machine learning
Neural architecture search

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-030-63836-8_24

Cite this

Pauletto, L., Amini, M. R., Babbar, R., & Winckler, N. (2020). Neural Architecture Search for Extreme Multi-label Text Classification. In H. Yang, K. Pasupa, A. C.-S. Leung, J. T. Kwok, J. H. Chan, & I. King (Eds.), Neural Information Processing - 27th International Conference, ICONIP 2020, Proceedings (pp. 282-293). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12534 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-63836-8_24

Neural Architecture Search for Extreme Multi-label Text Classification. / Pauletto, Loïc; Amini, Massih Reza; Babbar, Rohit et al.
Neural Information Processing - 27th International Conference, ICONIP 2020, Proceedings. ed. / Haiqin Yang; Kitsuchart Pasupa; Andrew Chi-Sing Leung; James T. Kwok; Jonathan H. Chan; Irwin King. Springer Science and Business Media Deutschland GmbH, 2020. p. 282-293 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12534 LNCS).

Research output: Chapter or section in a book/report/conference proceeding › Chapter in a published conference proceeding

Pauletto, L, Amini, MR, Babbar, R & Winckler, N 2020, Neural Architecture Search for Extreme Multi-label Text Classification. in H Yang, K Pasupa, AC-S Leung, JT Kwok, JH Chan & I King (eds), Neural Information Processing - 27th International Conference, ICONIP 2020, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12534 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 282-293, 27th International Conference on Neural Information Processing, ICONIP 2020, Bangkok, Thailand, 18/11/20. https://doi.org/10.1007/978-3-030-63836-8_24

Pauletto L, Amini MR, Babbar R, Winckler N. Neural Architecture Search for Extreme Multi-label Text Classification. In Yang H, Pasupa K, Leung ACS, Kwok JT, Chan JH, King I, editors, Neural Information Processing - 27th International Conference, ICONIP 2020, Proceedings. Springer Science and Business Media Deutschland GmbH. 2020. p. 282-293. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-63836-8_24

Pauletto, Loïc ; Amini, Massih Reza ; Babbar, Rohit et al. / Neural Architecture Search for Extreme Multi-label Text Classification. Neural Information Processing - 27th International Conference, ICONIP 2020, Proceedings. editor / Haiqin Yang ; Kitsuchart Pasupa ; Andrew Chi-Sing Leung ; James T. Kwok ; Jonathan H. Chan ; Irwin King. Springer Science and Business Media Deutschland GmbH, 2020. pp. 282-293 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{d20bb8e2ebdb4af2b761c496dd8cf1a4,

title = "Neural Architecture Search for Extreme Multi-label Text Classification",

abstract = "Extreme classification and Neural Architecture Search (NAS) are research topics which have recently gained a lot of interest. While the former has been mainly motivated and applied in e-commerce and Natural Language Processing (NLP) applications, the NAS approach has been applied to a small variety of tasks, mainly in image processing. In this study, we extend the scope of NAS to the task of extreme multilabel classification (XMC). We propose a neuro-evolution approach, which was found to be the most suitable for a variety of tasks. Our NAS method automatically finds architectures that give competitive results with respect to the state of the art (and superior to other methods) with faster convergence. In addition, we perform analysis of the weights of the architecture blocks to provide insight into the importance of different operations that have been selected by the method.",

keywords = "Evolutionary algorithms, Extreme multi-label text classification, Machine learning, Neural architecture search",

author = "Lo{\"i}c Pauletto and Amini, {Massih Reza} and Rohit Babbar and Nicolas Winckler",

note = "Publisher Copyright: {\textcopyright} 2020, Springer Nature Switzerland AG.; 27th International Conference on Neural Information Processing, ICONIP 2020 ; Conference date: 18-11-2020 Through 22-11-2020",

year = "2020",

month = nov,

day = "19",

doi = "10.1007/978-3-030-63836-8_24",

language = "English",

isbn = "9783030638351",

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

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "282--293",

editor = "Haiqin Yang and Kitsuchart Pasupa and Leung, {Andrew Chi-Sing} and Kwok, {James T.} and Chan, {Jonathan H.} and Irwin King",

booktitle = "Neural Information Processing - 27th International Conference, ICONIP 2020, Proceedings",

address = "Germany",

}

TY - GEN

T1 - Neural Architecture Search for Extreme Multi-label Text Classification

AU - Pauletto, Loïc

AU - Amini, Massih Reza

AU - Babbar, Rohit

AU - Winckler, Nicolas

PY - 2020/11/19

Y1 - 2020/11/19

N2 - Extreme classification and Neural Architecture Search (NAS) are research topics which have recently gained a lot of interest. While the former has been mainly motivated and applied in e-commerce and Natural Language Processing (NLP) applications, the NAS approach has been applied to a small variety of tasks, mainly in image processing. In this study, we extend the scope of NAS to the task of extreme multilabel classification (XMC). We propose a neuro-evolution approach, which was found to be the most suitable for a variety of tasks. Our NAS method automatically finds architectures that give competitive results with respect to the state of the art (and superior to other methods) with faster convergence. In addition, we perform analysis of the weights of the architecture blocks to provide insight into the importance of different operations that have been selected by the method.

AB - Extreme classification and Neural Architecture Search (NAS) are research topics which have recently gained a lot of interest. While the former has been mainly motivated and applied in e-commerce and Natural Language Processing (NLP) applications, the NAS approach has been applied to a small variety of tasks, mainly in image processing. In this study, we extend the scope of NAS to the task of extreme multilabel classification (XMC). We propose a neuro-evolution approach, which was found to be the most suitable for a variety of tasks. Our NAS method automatically finds architectures that give competitive results with respect to the state of the art (and superior to other methods) with faster convergence. In addition, we perform analysis of the weights of the architecture blocks to provide insight into the importance of different operations that have been selected by the method.

KW - Evolutionary algorithms

KW - Extreme multi-label text classification

KW - Machine learning

KW - Neural architecture search

UR - http://www.scopus.com/inward/record.url?scp=85097438471&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-63836-8_24

DO - 10.1007/978-3-030-63836-8_24

M3 - Chapter in a published conference proceeding

AN - SCOPUS:85097438471

SN - 9783030638351

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

SP - 282

EP - 293

BT - Neural Information Processing - 27th International Conference, ICONIP 2020, Proceedings

A2 - Yang, Haiqin

A2 - Pasupa, Kitsuchart

A2 - Leung, Andrew Chi-Sing

A2 - Kwok, James T.

A2 - Chan, Jonathan H.

A2 - King, Irwin

PB - Springer Science and Business Media Deutschland GmbH

T2 - 27th International Conference on Neural Information Processing, ICONIP 2020

Y2 - 18 November 2020 through 22 November 2020

ER -

Neural Architecture Search for Extreme Multi-label Text Classification

Abstract

Publication series

Conference

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this