Fraunhofer-Gesellschaft

Publica

Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Applications of lp-norms and their smooth approximations for gradient based learning vector quantization

 
: Lange, M.; Zühlke, D.; Holz, O.; Villmann, T.

Verleysen, M. ; Katholieke Universiteit, Leuven:
22nd European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2014. Proceedings : Bruges, Belgium, April 23 - 25, 2014
Louvain-la-Neuve: Ciaco, 2014
ISBN: 978-2-87419-095-7
pp.271-276
European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN) <22, 2014, Bruges>
English
Conference Paper
Fraunhofer IAIS ()
Fraunhofer ITEM ()

Abstract
Learning vector quantization applying non-standard metrics became quite popular for classification performance improvement compared to standard approaches using the Euclidean distance. Kernel metrics and quadratic forms belong to the most promising approaches. In this paper we consider Minkowski distances (lp-norms). In particular, l1-norms are known to be robust against noise in data, such that, if this structural knowledge is available in advance about the data, this norm should be utilized. However, application in gradient based learning algorithms based on distance evaluations need to calculate the respective derivatives. Because lp-distance formulas contain the absolute approximations thereof are required. We consider in this paper several approaches for smooth consistent approximations for numerical evaluations and demonstrate the applicability for exemplary real world applications.

: http://publica.fraunhofer.de/documents/N-441942.html