Unimodal regularisation based on beta distribution for deep ordinal regression

Authors

Víctor Manuel Vargas-Yun

Pedro Antonio Gutiérrez

César Hervás-Martínez

Published in Journal

Pattern Recognition

COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

Impact Factor JCR 2022

JCR Ranking

22 / 145

Position

ISSN 0031-3203

Vol. 122

Pages 108310

Abstract

Currently, the use of deep learning for solving ordinal classification problems, where categories follow a natural order, has not received much attention. In this paper, we propose an unimodal regularisation based on the beta distribution applied to the cross-entropy loss. This regularisation encourages the distribution of the labels to be a soft unimodal distribution, more appropriate for ordinal problems. Given that the beta distribution has two parameters that must be adjusted, a method to automatically determine them is proposed. The regularised loss function is used to train a deep neural network model with an ordinal scheme in the output layer. The results obtained are statistically analysed and show that the combination of these methods increases the performance in ordinal problems. Moreover, the proposed beta distribution performs better than other distributions proposed in previous works, achieving also a reduced computational cost.

Keywords

beta distribution convolutional network ordinal regression stick-breaking unimodal distribution

BibTex Citation

@article{Vargas2022Unimodal,
	author = {Vargas-Yun, V{\' i}ctor Manuel and Guti{\' e}rrez, Pedro Antonio and Herv{\' a}s-Mart{\' i}nez, C{\' e}sar},
	journal = {Pattern Recognition},
	doi = {10.1016/j.patcog.2021.108310},
	year = {2022},
	pages = {108310},
	title = {Unimodal regularisation based on beta distribution for deep ordinal regression},
	url = {doi.org/10.1016/j.patcog.2021.108310},
	howpublished = {doi.org/10.1016/j.patcog.2021.108310},
	volume = {122},
}

BibTex Unicode Citation

@article{Vargas2022Unimodal,
	author = {Vargas-Yun, Víctor Manuel and Gutiérrez, Pedro Antonio and Hervás-Martínez, César},
	journal = {Pattern Recognition},
	doi = {10.1016/j.patcog.2021.108310},
	year = {2022},
	pages = {108310},
	title = {Unimodal regularisation based on beta distribution for deep ordinal regression},
	url = {doi.org/10.1016/j.patcog.2021.108310},
	howpublished = {doi.org/10.1016/j.patcog.2021.108310},
	volume = {122},
}

APA Citation

Vargas-Yun, V. M., Gutiérrez, P. A., & Hervás-Martínez, C. (2022). Unimodal regularisation based on beta distribution for deep ordinal regression. Pattern Recognition, 122, 108310. https://doi.org/10.1016/j.patcog.2021.108310

RIS Citation

TY  - JOUR
AU  - Vargas-Yun, Víctor Manuel
AU  - Gutiérrez, Pedro Antonio
AU  - Hervás-Martínez, César
DA  - 2022///
PY  - 2022
DO  - 10.1016/j.patcog.2021.108310
ID  - temp_id_786533331034
SP  - 108310
T2  - Pattern Recognition
TI  - Unimodal regularisation based on beta distribution for deep ordinal re
gression
UR  - doi.org/10.1016/j.patcog.2021.108310
VL  - 122
ER  -

CV Citation

V.M. Vargas-Yun, P.A. Gutiérrez, C. Hervás-Martínez (1/3). "Unimodal regularisation based on beta distribution for deep ordinal regression". Pattern Recognition,  Vol. 122, pp. 108310, 2022. (Q1, IF: 8.0).