Multitask learning using regularized multiple kernel learning

Mehmet Gönen, Melih Kandemir, Samuel Kaski

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Scopus citations


Empirical success of kernel-based learning algorithms is very much dependent on the kernel function used. Instead of using a single fixed kernel function, multiple kernel learning (MKL) algorithms learn a combination of different kernel functions in order to obtain a similarity measure that better matches the underlying problem. We study multitask learning (MKL) problems and formulate a novel MTL algorithm that trains coupled but nonidentical MKL models across the tasks. The proposed algorithm is especially useful for tasks that have different input and/or output space characteristics and is computationally very efficient. Empirical results on three data sets validate the generalization performance and the efficiency of our approach.

Original languageEnglish (US)
Title of host publicationNeural Information Processing - 18th International Conference, ICONIP 2011, Proceedings
Number of pages10
EditionPART 2
StatePublished - 2011
Externally publishedYes
Event18th International Conference on Neural Information Processing, ICONIP 2011 - Shanghai, China
Duration: Nov 13 2011Nov 17 2011

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 2
Volume7063 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Other18th International Conference on Neural Information Processing, ICONIP 2011


  • kernel machines
  • multilabel learning
  • multiple kernel learning
  • multitask learning
  • support vector machines

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)


Dive into the research topics of 'Multitask learning using regularized multiple kernel learning'. Together they form a unique fingerprint.

Cite this