Unimodal regularisation based on beta distribution for deep ordinal regression

V. Vargas , P. Gutiérrez, C. Hervás-Martínez

Pattern Recognition, Vol. 122, pp. 1-10, 2022 Indexed in JCR. Impact factor: 8, Position: 25/145 (Q1) in COMPUTER SCIENCE, ARTIFICIAL INTELLIGENCE

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.

Cite this publication

BibTex

@article{vargas2022unimodal,
    author = {Víctor Manuel Vargas and Pedro Antonio Gutiérrez and César Hervás-Martínez},
    title = {Unimodal regularisation based on beta distribution for deep ordinal regression},
    journal = {Pattern Recognition},
    year = {2022},
    volume = {122},
    number = {0},
    pages = {1--10},
    doi = {10.1016/j.patcog.2021.108310}
}

APA

Vargas, V., Gutiérrez, P., Hervás-Martínez, C. (2022). Unimodal regularisation based on beta distribution for deep ordinal regression. Pattern Recognition, 122(0), 1-10.

V.M. Vargas (CA), P.A. Gutiérrez, C. Hervás-Martínez, (1/3) "Unimodal regularisation based on beta distribution for deep ordinal regression", Pattern Recognition, Vol. 122(0), pp. 1-10, 2022. (Q1, IF: 8)

RIS

TY  - JOUR
T1  - Unimodal regularisation based on beta distribution for deep ordinal regression
AU  - Vargas, Víctor Manuel
AU  - Gutiérrez, Pedro Antonio
AU  - Hervás-Martínez, César
JO  - Pattern Recognition
VL  - 122
IS  - 0
SP  - 1
EP  - 10
PY  - 2022
DO  - 10.1016/j.patcog.2021.108310
ER  -