Learning-Based Confidence Estimation for Multi-modal Classifier Fusion

Uzair Nadeem; Mohammed Bennamoun; Ferdous Sohel; Roberto Togneri

doi:10.1007/978-3-030-36711-4_26

Learning-Based Confidence Estimation for Multi-modal Classifier Fusion

Uzair Nadeem, Mohammed Bennamoun, Ferdous Sohel, Roberto Togneri

Research output: Chapter in Book/Conference paper › Conference paper › peer-review

5 Citations (Scopus)

Abstract

We propose a novel confidence estimation method for predictions from a multi-class classifier. Unlike existing methods, we learn a confidence-estimator on the basis of a held-out set from the training data. The predicted confidence values by the proposed system are used to improve the accuracy of multi-modal emotion and sentiment classification. The scores of different classes from the individual modalities are superposed on the basis of confidence values. Experimental results demonstrate that the accuracy of the proposed confidence based fusion method is significantly superior to that of the classifier trained on any modality separately, and achieves superior performance compared to other fusion methods.

Original language	English
Title of host publication	Neural Information Processing - 26th International Conference, ICONIP 2019, Proceedings
Subtitle of host publication	part II
Editors	Tom Gedeon, Kok Wai Wong, Minho Lee
Place of Publication	Australia
Publisher	Springer
Pages	299-312
Number of pages	14
ISBN (Print)	9783030367107
DOIs	https://doi.org/10.1007/978-3-030-36711-4_26
Publication status	Published - 15 Dec 2019
Event	26th International Conference on Neural Information Processing - Sydney, Australia Duration: 12 Dec 2019 → 15 Dec 2019 http://ajiips.com.au/iconip2019/

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11954 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	26th International Conference on Neural Information Processing
Abbreviated title	ICONIP 2019
Country/Territory	Australia
City	Sydney
Period	12/12/19 → 15/12/19
Internet address	http://ajiips.com.au/iconip2019/

Access to Document

10.1007/978-3-030-36711-4_26

Cite this

Nadeem, U., Bennamoun, M., Sohel, F., & Togneri, R. (2019). Learning-Based Confidence Estimation for Multi-modal Classifier Fusion. In T. Gedeon, K. W. Wong, & M. Lee (Eds.), Neural Information Processing - 26th International Conference, ICONIP 2019, Proceedings: part II (pp. 299-312). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11954 LNCS). Springer. https://doi.org/10.1007/978-3-030-36711-4_26

Nadeem, Uzair ; Bennamoun, Mohammed ; Sohel, Ferdous et al. / Learning-Based Confidence Estimation for Multi-modal Classifier Fusion. Neural Information Processing - 26th International Conference, ICONIP 2019, Proceedings: part II. editor / Tom Gedeon ; Kok Wai Wong ; Minho Lee. Australia : Springer, 2019. pp. 299-312 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{db5b0f6490d8412493468edf0e5de6a8,

title = "Learning-Based Confidence Estimation for Multi-modal Classifier Fusion",

abstract = "We propose a novel confidence estimation method for predictions from a multi-class classifier. Unlike existing methods, we learn a confidence-estimator on the basis of a held-out set from the training data. The predicted confidence values by the proposed system are used to improve the accuracy of multi-modal emotion and sentiment classification. The scores of different classes from the individual modalities are superposed on the basis of confidence values. Experimental results demonstrate that the accuracy of the proposed confidence based fusion method is significantly superior to that of the classifier trained on any modality separately, and achieves superior performance compared to other fusion methods.",

keywords = "Confidence estimation, Emotion classification, Multi-modal fusion, Sentiment analysis",

author = "Uzair Nadeem and Mohammed Bennamoun and Ferdous Sohel and Roberto Togneri",

year = "2019",

month = dec,

day = "15",

doi = "10.1007/978-3-030-36711-4_26",

language = "English",

isbn = "9783030367107",

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

publisher = "Springer",

pages = "299--312",

editor = "Tom Gedeon and Wong, {Kok Wai} and Minho Lee",

booktitle = "Neural Information Processing - 26th International Conference, ICONIP 2019, Proceedings",

address = "Netherlands",

note = "26th International Conference on Neural Information Processing, ICONIP 2019 ; Conference date: 12-12-2019 Through 15-12-2019",

url = "http://ajiips.com.au/iconip2019/",

}

Nadeem, U, Bennamoun, M, Sohel, F & Togneri, R 2019, Learning-Based Confidence Estimation for Multi-modal Classifier Fusion. in T Gedeon, KW Wong & M Lee (eds), Neural Information Processing - 26th International Conference, ICONIP 2019, Proceedings: part II. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11954 LNCS, Springer, Australia, pp. 299-312, 26th International Conference on Neural Information Processing, Sydney, New South Wales, Australia, 12/12/19. https://doi.org/10.1007/978-3-030-36711-4_26

Learning-Based Confidence Estimation for Multi-modal Classifier Fusion. / Nadeem, Uzair; Bennamoun, Mohammed; Sohel, Ferdous et al.
Neural Information Processing - 26th International Conference, ICONIP 2019, Proceedings: part II. ed. / Tom Gedeon; Kok Wai Wong; Minho Lee. Australia: Springer, 2019. p. 299-312 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11954 LNCS).

Research output: Chapter in Book/Conference paper › Conference paper › peer-review

TY - GEN

T1 - Learning-Based Confidence Estimation for Multi-modal Classifier Fusion

AU - Nadeem, Uzair

AU - Bennamoun, Mohammed

AU - Sohel, Ferdous

AU - Togneri, Roberto

PY - 2019/12/15

Y1 - 2019/12/15

N2 - We propose a novel confidence estimation method for predictions from a multi-class classifier. Unlike existing methods, we learn a confidence-estimator on the basis of a held-out set from the training data. The predicted confidence values by the proposed system are used to improve the accuracy of multi-modal emotion and sentiment classification. The scores of different classes from the individual modalities are superposed on the basis of confidence values. Experimental results demonstrate that the accuracy of the proposed confidence based fusion method is significantly superior to that of the classifier trained on any modality separately, and achieves superior performance compared to other fusion methods.

AB - We propose a novel confidence estimation method for predictions from a multi-class classifier. Unlike existing methods, we learn a confidence-estimator on the basis of a held-out set from the training data. The predicted confidence values by the proposed system are used to improve the accuracy of multi-modal emotion and sentiment classification. The scores of different classes from the individual modalities are superposed on the basis of confidence values. Experimental results demonstrate that the accuracy of the proposed confidence based fusion method is significantly superior to that of the classifier trained on any modality separately, and achieves superior performance compared to other fusion methods.

KW - Confidence estimation

KW - Emotion classification

KW - Multi-modal fusion

KW - Sentiment analysis

UR - http://www.scopus.com/inward/record.url?scp=85076918806&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-36711-4_26

DO - 10.1007/978-3-030-36711-4_26

M3 - Conference paper

AN - SCOPUS:85076918806

SN - 9783030367107

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

SP - 299

EP - 312

BT - Neural Information Processing - 26th International Conference, ICONIP 2019, Proceedings

A2 - Gedeon, Tom

A2 - Wong, Kok Wai

A2 - Lee, Minho

PB - Springer

CY - Australia

T2 - 26th International Conference on Neural Information Processing

Y2 - 12 December 2019 through 15 December 2019

ER -

Nadeem U, Bennamoun M, Sohel F, Togneri R. Learning-Based Confidence Estimation for Multi-modal Classifier Fusion. In Gedeon T, Wong KW, Lee M, editors, Neural Information Processing - 26th International Conference, ICONIP 2019, Proceedings: part II. Australia: Springer. 2019. p. 299-312. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-36711-4_26

Learning-Based Confidence Estimation for Multi-modal Classifier Fusion

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this