J. Wehrmann, R. Cerri, and R. Barros. Proceedings of the 35th International Conference on Machine Learning, volume 80 of Proceedings of Machine Learning Research, page 5075--5084. Stockholmsmässan, Stockholm Sweden, PMLR, (10--15 Jul 2018)
Abstract
One of the most challenging machine learning problems is a particular case of data classification in which classes are hierarchically structured and objects can be assigned to multiple paths of the class hierarchy at the same time. This task is known as hierarchical multi-label classification (HMC), with applications in text classification, image annotation, and in bioinformatics problems such as protein function prediction. In this paper, we propose novel neural network architectures for HMC called HMCN, capable of simultaneously optimizing local and global loss functions for discovering local hierarchical class-relationships and global information from the entire class hierarchy while penalizing hierarchical violations. We evaluate its performance in 21 datasets from four distinct domains, and we compare it against the current HMC state-of-the-art approaches. Results show that HMCN substantially outperforms all baselines with statistical significance, arising as the novel state-of-the-art for HMC.
%0 Conference Paper
%1 wehrmann2018hierarchical
%A Wehrmann, Jonatas
%A Cerri, Ricardo
%A Barros, Rodrigo
%B Proceedings of the 35th International Conference on Machine Learning
%C Stockholmsmässan, Stockholm Sweden
%D 2018
%E Dy, Jennifer
%E Krause, Andreas
%I PMLR
%K classification hierarchical multilabel
%P 5075--5084
%T Hierarchical Multi-Label Classification Networks
%U http://proceedings.mlr.press/v80/wehrmann18a.html
%V 80
%X One of the most challenging machine learning problems is a particular case of data classification in which classes are hierarchically structured and objects can be assigned to multiple paths of the class hierarchy at the same time. This task is known as hierarchical multi-label classification (HMC), with applications in text classification, image annotation, and in bioinformatics problems such as protein function prediction. In this paper, we propose novel neural network architectures for HMC called HMCN, capable of simultaneously optimizing local and global loss functions for discovering local hierarchical class-relationships and global information from the entire class hierarchy while penalizing hierarchical violations. We evaluate its performance in 21 datasets from four distinct domains, and we compare it against the current HMC state-of-the-art approaches. Results show that HMCN substantially outperforms all baselines with statistical significance, arising as the novel state-of-the-art for HMC.
@inproceedings{wehrmann2018hierarchical,
abstract = {One of the most challenging machine learning problems is a particular case of data classification in which classes are hierarchically structured and objects can be assigned to multiple paths of the class hierarchy at the same time. This task is known as hierarchical multi-label classification (HMC), with applications in text classification, image annotation, and in bioinformatics problems such as protein function prediction. In this paper, we propose novel neural network architectures for HMC called HMCN, capable of simultaneously optimizing local and global loss functions for discovering local hierarchical class-relationships and global information from the entire class hierarchy while penalizing hierarchical violations. We evaluate its performance in 21 datasets from four distinct domains, and we compare it against the current HMC state-of-the-art approaches. Results show that HMCN substantially outperforms all baselines with statistical significance, arising as the novel state-of-the-art for HMC.},
added-at = {2018-12-10T11:47:59.000+0100},
address = {Stockholmsmässan, Stockholm Sweden},
author = {Wehrmann, Jonatas and Cerri, Ricardo and Barros, Rodrigo},
biburl = {https://www.bibsonomy.org/bibtex/2278c1be0567cc39dc7affd400e0121b4/thoni},
booktitle = {Proceedings of the 35th International Conference on Machine Learning},
description = {Hierarchical Multi-Label Classification Networks},
editor = {Dy, Jennifer and Krause, Andreas},
interhash = {23320ea2aa9873647bac6373fcc5a18f},
intrahash = {278c1be0567cc39dc7affd400e0121b4},
keywords = {classification hierarchical multilabel},
month = {10--15 Jul},
pages = {5075--5084},
pdf = {http://proceedings.mlr.press/v80/wehrmann18a/wehrmann18a.pdf},
publisher = {PMLR},
series = {Proceedings of Machine Learning Research},
timestamp = {2018-12-10T11:47:59.000+0100},
title = {Hierarchical Multi-Label Classification Networks},
url = {http://proceedings.mlr.press/v80/wehrmann18a.html},
volume = 80,
year = 2018
}