Multitask learning using regularized multiple kernel learning

Mehmet Gönen, Melih Kandemir, Samuel Kaski

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

3 Scopus citations

Abstract

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
Pages500-509
Number of pages10
EditionPART 2
DOIs
StatePublished - Nov 28 2011
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

Other

Other18th International Conference on Neural Information Processing, ICONIP 2011
CountryChina
CityShanghai
Period11/13/1111/17/11

Keywords

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

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

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

  • Cite this

    Gönen, M., Kandemir, M., & Kaski, S. (2011). Multitask learning using regularized multiple kernel learning. In Neural Information Processing - 18th International Conference, ICONIP 2011, Proceedings (PART 2 ed., pp. 500-509). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7063 LNCS, No. PART 2). https://doi.org/10.1007/978-3-642-24958-7_58