Maximizing subset accuracy with recurrent neural networks in multi-label classification

Jinseok Nam, Eneldo Loza Mencía, Hyun Woo Kim, Johannes Fürnkranz

Research output: Contribution to journalConference article

13 Citations (Scopus)

Abstract

Multi-label classification is the task of predicting a set of labels for a given input instance. Classifier chains are a state-of-the-art method for tackling such problems, which essentially converts this problem into a sequential prediction problem, where the labels are first ordered in an arbitrary fashion, and the task is to predict a sequence of binary values for these labels. In this paper, we replace classifier chains with recurrent neural networks, a sequence-to-sequence prediction algorithm which has recently been successfully applied to sequential prediction tasks in many domains. The key advantage of this approach is that it allows to focus on the prediction of the positive labels only, a much smaller set than the full set of possible labels. Moreover, parameter sharing across all classifiers allows to better exploit information of previous decisions. As both, classifier chains and recurrent neural networks depend on a fixed ordering of the labels, which is typically not part of a multi-label problem specification, we also compare different ways of ordering the label set, and give some recommendations on suitable ordering strategies.

Original languageEnglish
Pages (from-to)5414-5424
Number of pages11
JournalAdvances in Neural Information Processing Systems
Volume2017-December
Publication statusPublished - 2017 Jan 1
Externally publishedYes
Event31st Annual Conference on Neural Information Processing Systems, NIPS 2017 - Long Beach, United States
Duration: 2017 Dec 42017 Dec 9

Fingerprint

Recurrent neural networks
Labels
Classifiers
Specifications

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Information Systems
  • Signal Processing

Cite this

Maximizing subset accuracy with recurrent neural networks in multi-label classification. / Nam, Jinseok; Mencía, Eneldo Loza; Kim, Hyun Woo; Fürnkranz, Johannes.

In: Advances in Neural Information Processing Systems, Vol. 2017-December, 01.01.2017, p. 5414-5424.

Research output: Contribution to journalConference article

Nam, Jinseok ; Mencía, Eneldo Loza ; Kim, Hyun Woo ; Fürnkranz, Johannes. / Maximizing subset accuracy with recurrent neural networks in multi-label classification. In: Advances in Neural Information Processing Systems. 2017 ; Vol. 2017-December. pp. 5414-5424.
@article{995099d315c34158a9c0af073d11dc0a,
title = "Maximizing subset accuracy with recurrent neural networks in multi-label classification",
abstract = "Multi-label classification is the task of predicting a set of labels for a given input instance. Classifier chains are a state-of-the-art method for tackling such problems, which essentially converts this problem into a sequential prediction problem, where the labels are first ordered in an arbitrary fashion, and the task is to predict a sequence of binary values for these labels. In this paper, we replace classifier chains with recurrent neural networks, a sequence-to-sequence prediction algorithm which has recently been successfully applied to sequential prediction tasks in many domains. The key advantage of this approach is that it allows to focus on the prediction of the positive labels only, a much smaller set than the full set of possible labels. Moreover, parameter sharing across all classifiers allows to better exploit information of previous decisions. As both, classifier chains and recurrent neural networks depend on a fixed ordering of the labels, which is typically not part of a multi-label problem specification, we also compare different ways of ordering the label set, and give some recommendations on suitable ordering strategies.",
author = "Jinseok Nam and Menc{\'i}a, {Eneldo Loza} and Kim, {Hyun Woo} and Johannes F{\"u}rnkranz",
year = "2017",
month = "1",
day = "1",
language = "English",
volume = "2017-December",
pages = "5414--5424",
journal = "Advances in Neural Information Processing Systems",
issn = "1049-5258",

}

TY - JOUR

T1 - Maximizing subset accuracy with recurrent neural networks in multi-label classification

AU - Nam, Jinseok

AU - Mencía, Eneldo Loza

AU - Kim, Hyun Woo

AU - Fürnkranz, Johannes

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Multi-label classification is the task of predicting a set of labels for a given input instance. Classifier chains are a state-of-the-art method for tackling such problems, which essentially converts this problem into a sequential prediction problem, where the labels are first ordered in an arbitrary fashion, and the task is to predict a sequence of binary values for these labels. In this paper, we replace classifier chains with recurrent neural networks, a sequence-to-sequence prediction algorithm which has recently been successfully applied to sequential prediction tasks in many domains. The key advantage of this approach is that it allows to focus on the prediction of the positive labels only, a much smaller set than the full set of possible labels. Moreover, parameter sharing across all classifiers allows to better exploit information of previous decisions. As both, classifier chains and recurrent neural networks depend on a fixed ordering of the labels, which is typically not part of a multi-label problem specification, we also compare different ways of ordering the label set, and give some recommendations on suitable ordering strategies.

AB - Multi-label classification is the task of predicting a set of labels for a given input instance. Classifier chains are a state-of-the-art method for tackling such problems, which essentially converts this problem into a sequential prediction problem, where the labels are first ordered in an arbitrary fashion, and the task is to predict a sequence of binary values for these labels. In this paper, we replace classifier chains with recurrent neural networks, a sequence-to-sequence prediction algorithm which has recently been successfully applied to sequential prediction tasks in many domains. The key advantage of this approach is that it allows to focus on the prediction of the positive labels only, a much smaller set than the full set of possible labels. Moreover, parameter sharing across all classifiers allows to better exploit information of previous decisions. As both, classifier chains and recurrent neural networks depend on a fixed ordering of the labels, which is typically not part of a multi-label problem specification, we also compare different ways of ordering the label set, and give some recommendations on suitable ordering strategies.

UR - http://www.scopus.com/inward/record.url?scp=85047012313&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85047012313&partnerID=8YFLogxK

M3 - Conference article

VL - 2017-December

SP - 5414

EP - 5424

JO - Advances in Neural Information Processing Systems

JF - Advances in Neural Information Processing Systems

SN - 1049-5258

ER -