Label-noise-tolerant classification for streaming data

Benoit Frenay, Barbara Hammer

Résultats de recherche: RechercheArticle dans les actes d'une conférence/un colloque

Résumé

Label noise-tolerant machine learning techniques address datasets which are affected by mislabelling of the instances. Since labelling quality is a severe issue in particular for large or streaming data sets, this setting becomes more and more relevant in the context of life-long learning, big data and crowd sourcing. In this contribution, we extend a powerful online learning method, soft robust learning vector quantisation, by a probabilistic model for noise tolerance, which is applicable for streaming data, including label-noise drift. The superiority of the technique is demonstrated in several benchmark problems.

langueAnglais
titre2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings
EditeurInstitute of Electrical and Electronics Engineers Inc.
Pages1748-1755
Nombre de pages8
Volume2017-May
ISBN (Electronique)9781509061815
Les DOIs
étatPublié - 30 juin 2017
Evénement2017 International Joint Conference on Neural Networks, IJCNN 2017 - Anchorage, États-Unis
Durée: 14 mai 201719 mai 2017

Une conférence

Une conférence2017 International Joint Conference on Neural Networks, IJCNN 2017
PaysÉtats-Unis
La villeAnchorage
période14/05/1719/05/17

Empreinte digitale

Labels
Vector quantization
Labeling
Learning systems
Big data
Statistical Models

Citer ceci

Frenay, B., & Hammer, B. (2017). Label-noise-tolerant classification for streaming data. Dans 2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings (Vol 2017-May, p. 1748-1755). [7966062] Institute of Electrical and Electronics Engineers Inc.. DOI: 10.1109/IJCNN.2017.7966062
Frenay, Benoit ; Hammer, Barbara. / Label-noise-tolerant classification for streaming data. 2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings. Vol 2017-May Institute of Electrical and Electronics Engineers Inc., 2017. p. 1748-1755
@inbook{6168a89c7b6647e88994dfe6cf1ef287,
title = "Label-noise-tolerant classification for streaming data",
abstract = "Label noise-tolerant machine learning techniques address datasets which are affected by mislabelling of the instances. Since labelling quality is a severe issue in particular for large or streaming data sets, this setting becomes more and more relevant in the context of life-long learning, big data and crowd sourcing. In this contribution, we extend a powerful online learning method, soft robust learning vector quantisation, by a probabilistic model for noise tolerance, which is applicable for streaming data, including label-noise drift. The superiority of the technique is demonstrated in several benchmark problems.",
author = "Benoit Frenay and Barbara Hammer",
year = "2017",
month = "6",
doi = "10.1109/IJCNN.2017.7966062",
volume = "2017-May",
pages = "1748--1755",
booktitle = "2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

Frenay, B & Hammer, B 2017, Label-noise-tolerant classification for streaming data. Dans 2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings. VOL. 2017-May, 7966062, Institute of Electrical and Electronics Engineers Inc., p. 1748-1755, 2017 International Joint Conference on Neural Networks, IJCNN 2017, Anchorage, États-Unis, 14/05/17. DOI: 10.1109/IJCNN.2017.7966062

Label-noise-tolerant classification for streaming data. / Frenay, Benoit; Hammer, Barbara.

2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings. Vol 2017-May Institute of Electrical and Electronics Engineers Inc., 2017. p. 1748-1755 7966062.

Résultats de recherche: RechercheArticle dans les actes d'une conférence/un colloque

TY - CHAP

T1 - Label-noise-tolerant classification for streaming data

AU - Frenay,Benoit

AU - Hammer,Barbara

PY - 2017/6/30

Y1 - 2017/6/30

N2 - Label noise-tolerant machine learning techniques address datasets which are affected by mislabelling of the instances. Since labelling quality is a severe issue in particular for large or streaming data sets, this setting becomes more and more relevant in the context of life-long learning, big data and crowd sourcing. In this contribution, we extend a powerful online learning method, soft robust learning vector quantisation, by a probabilistic model for noise tolerance, which is applicable for streaming data, including label-noise drift. The superiority of the technique is demonstrated in several benchmark problems.

AB - Label noise-tolerant machine learning techniques address datasets which are affected by mislabelling of the instances. Since labelling quality is a severe issue in particular for large or streaming data sets, this setting becomes more and more relevant in the context of life-long learning, big data and crowd sourcing. In this contribution, we extend a powerful online learning method, soft robust learning vector quantisation, by a probabilistic model for noise tolerance, which is applicable for streaming data, including label-noise drift. The superiority of the technique is demonstrated in several benchmark problems.

UR - http://www.scopus.com/inward/record.url?scp=85030979635&partnerID=8YFLogxK

U2 - 10.1109/IJCNN.2017.7966062

DO - 10.1109/IJCNN.2017.7966062

M3 - Conference contribution

VL - 2017-May

SP - 1748

EP - 1755

BT - 2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Frenay B, Hammer B. Label-noise-tolerant classification for streaming data. Dans 2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings. Vol 2017-May. Institute of Electrical and Electronics Engineers Inc.2017. p. 1748-1755. 7966062. Disponible à, DOI: 10.1109/IJCNN.2017.7966062