Multiple kernel learning for brain-computer interfacing

Wojciech Samek, Alexander Binder, Klaus Robert Muller

Research output: Contribution to journalArticle

Abstract

Combining information from different sources is a common way to improve classification accuracy in Brain-Computer Interfacing (BCI). For instance, in small sample settings it is useful to integrate data from other subjects or sessions in order to improve the estimation quality of the spatial filters or the classifier. Since data from different subjects may show large variability, it is crucial to weight the contributions according to importance. Many multi-subject learning algorithms determine the optimal weighting in a separate step by using heuristics, however, without ensuring that the selected weights are optimal with respect to classification. In this work we apply Multiple Kernel Learning (MKL) to this problem. MKL has been widely used for feature fusion in computer vision and allows to simultaneously learn the classifier and the optimal weighting. We compare the MKL method to two baseline approaches and investigate the reasons for performance improvement.

Fingerprint

Brain
Classifiers
Learning
Learning algorithms
Computer vision
Fusion reactions
Weights and Measures

ASJC Scopus subject areas

  • Medicine(all)

Cite this

@article{17f8c227f0f74ce8a846a5f1bcd2ff3c,
title = "Multiple kernel learning for brain-computer interfacing",
abstract = "Combining information from different sources is a common way to improve classification accuracy in Brain-Computer Interfacing (BCI). For instance, in small sample settings it is useful to integrate data from other subjects or sessions in order to improve the estimation quality of the spatial filters or the classifier. Since data from different subjects may show large variability, it is crucial to weight the contributions according to importance. Many multi-subject learning algorithms determine the optimal weighting in a separate step by using heuristics, however, without ensuring that the selected weights are optimal with respect to classification. In this work we apply Multiple Kernel Learning (MKL) to this problem. MKL has been widely used for feature fusion in computer vision and allows to simultaneously learn the classifier and the optimal weighting. We compare the MKL method to two baseline approaches and investigate the reasons for performance improvement.",
author = "Wojciech Samek and Alexander Binder and Muller, {Klaus Robert}",
year = "2013",
doi = "10.1109/EMBC.2013.6611181",
language = "English",
volume = "2013",
pages = "7048--7051",
journal = "The BMJ",
issn = "0730-6512",
publisher = "Kluwer Academic Publishers",

}

TY - JOUR

T1 - Multiple kernel learning for brain-computer interfacing

AU - Samek, Wojciech

AU - Binder, Alexander

AU - Muller, Klaus Robert

PY - 2013

Y1 - 2013

N2 - Combining information from different sources is a common way to improve classification accuracy in Brain-Computer Interfacing (BCI). For instance, in small sample settings it is useful to integrate data from other subjects or sessions in order to improve the estimation quality of the spatial filters or the classifier. Since data from different subjects may show large variability, it is crucial to weight the contributions according to importance. Many multi-subject learning algorithms determine the optimal weighting in a separate step by using heuristics, however, without ensuring that the selected weights are optimal with respect to classification. In this work we apply Multiple Kernel Learning (MKL) to this problem. MKL has been widely used for feature fusion in computer vision and allows to simultaneously learn the classifier and the optimal weighting. We compare the MKL method to two baseline approaches and investigate the reasons for performance improvement.

AB - Combining information from different sources is a common way to improve classification accuracy in Brain-Computer Interfacing (BCI). For instance, in small sample settings it is useful to integrate data from other subjects or sessions in order to improve the estimation quality of the spatial filters or the classifier. Since data from different subjects may show large variability, it is crucial to weight the contributions according to importance. Many multi-subject learning algorithms determine the optimal weighting in a separate step by using heuristics, however, without ensuring that the selected weights are optimal with respect to classification. In this work we apply Multiple Kernel Learning (MKL) to this problem. MKL has been widely used for feature fusion in computer vision and allows to simultaneously learn the classifier and the optimal weighting. We compare the MKL method to two baseline approaches and investigate the reasons for performance improvement.

UR - http://www.scopus.com/inward/record.url?scp=84941131675&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84941131675&partnerID=8YFLogxK

U2 - 10.1109/EMBC.2013.6611181

DO - 10.1109/EMBC.2013.6611181

M3 - Article

VL - 2013

SP - 7048

EP - 7051

JO - The BMJ

JF - The BMJ

SN - 0730-6512

ER -