Multi-task Learning via Non-sparse Multiple Kernel Learning

  • Wojciech Samek
  • Alexander Binder
  • Motoaki Kawanabe
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6854)

Abstract

In object classification tasks from digital photographs, multiple categories are considered for annotation. Some of these visual concepts may have semantic relations and can appear simultaneously in images. Although taxonomical relations and co-occurrence structures between object categories have been studied, it is not easy to use such information to enhance performance of object classification. In this paper, we propose a novel multi-task learning procedure which extracts useful information from the classifiers for the other categories. Our approach is based on non-sparse multiple kernel learning (MKL) which has been successfully applied to adaptive feature selection for image classification. Experimental results on PASCAL VOC 2009 data show the potential of our method.

Keywords

Image Annotation Multi-Task Learning Multiple Kernel Learning 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Argyriou, A., Evgeniou, T., Pontil, M.: Multi-task feature learning. In: Neural Inf. Proc. Sys. MIT Press, Cambridge (2007)Google Scholar
  2. 2.
    Everingham, M., Van Gool, L., Williams, C.K.I., Winn, J., Zisserman, A.: The PASCAL Visual Object Classes Challenge 2009 (VOC 2009) ResultsGoogle Scholar
  3. 3.
    Evgeniou, T., Micchelli, C.A., Pontil, M.: Learning multiple tasks with kernel methods. J. of Mach. Learn. Res. 6, 615–637 (2005)MathSciNetMATHGoogle Scholar
  4. 4.
    Kloft, M., Brefeld, U., Sonnenburg, S., Laskov, P., Müller, K.-R., Zien, A.: Efficient and accurate ℓp-norm mkl. In: Neural Inf. Proc. Sys. (2010)Google Scholar
  5. 5.
    Lampert, C., Blaschko, M.: A multiple kernel learning approach to joint multi-class object detection. In: Rigoll, G. (ed.) DAGM 2008. LNCS, vol. 5096, pp. 31–40. Springer, Heidelberg (2008)CrossRefGoogle Scholar
  6. 6.
    Lanckriet, G.R.G., Cristianini, N., Bartlett, P., El Ghaoui, L., Jordan, M.I.: Learning the kernel matrix with semidefinite programming. J. of Mach. Learn. Res., 27–72 (2004)Google Scholar
  7. 7.
    Lazebnik, S., Schmid, C., Ponce, J.: Beyond bags of features: Spatial pyramid matching for recognizing natural scene categories. In: IEEE Conf. on Comp. Vision & Pat. Rec. (2006)Google Scholar
  8. 8.
    Ming, K., Chai, A., Williams, C.K.I., Klanke, S., Vijayakumar, S.: Multi-task gaussian process learning of robot inverse dynamics. In: Neural Inf. Proc. Sys. (2008)Google Scholar
  9. 9.
    Sheldon, D.: Graphical multi-task learning. In: Neural Inf. Proc. Sys. Workshop on Strictured Input - Structured Output (2008)Google Scholar
  10. 10.
    van de Sande, K.E.A., Gevers, T., Snoek, C.G.M.: Evaluating color descriptors for object and scene recognition. IEEE Trans. Pat. Anal. & Mach. Intel. (2010)Google Scholar
  11. 11.
    Wu, W., Li, H., Hu, Y., Jin, R.: Multi-task learning: Multiple kernels for multiple tasks. Technical report, Microsoft Research (2010)Google Scholar
  12. 12.
    Yuan, X.-T., Yan, S.: Visual classification with multi-task joint sparse representation. In: IEEE Conf. on Comp. Vision & Pat. Rec., pp. 3493–3500 (2010)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2011

Authors and Affiliations

  • Wojciech Samek
    • 1
    • 2
  • Alexander Binder
    • 1
  • Motoaki Kawanabe
    • 2
    • 1
  1. 1.Technical University of BerlinBerlinGermany
  2. 2.Fraunhofer Institute FIRSTBerlinGermany

Personalised recommendations