On the Usefulness of Similarity Based Projection Spaces for Transfer Learning

Abstract : Similarity functions are widely used in many machine learning or pattern recognition tasks. We consider here a recent framework for binary classication, proposed by Balcan et al., allowing to learn in a potentially non geometrical space based on good similarity functions. This framework is a generalization of the notion of kernels used in support vector machines in the sense that allows ne to use similarity functions that do not need to be positive semi-de nite nor symmetric. The similarities are then used to de ne an xplicit projection space where a linear classi er with good generalization properties can be learned. In this paper, we propose to study experimentally the usefulness of similarity based projection spaces for transfer learning issues. More precisely, we consider the problem of domain adaptation where the distributions generating learning data and test data are somewhat different. We stand in the case where no information on the test labels is available. We show that a simple renormalization of a good similarity function taking into account the test data allows us to learn classifiers more performing on the target distribution for difficult adaptation problems. Moreover, this normalization always helps to improve the model when we try to regularize the similarity based projection space in order to move closer the two distributions. We provide experiments on a toy problem and on a real image annotation task.
Document type :
Conference papers
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-00628991
Contributor : Emilie Morvant <>
Submitted on : Tuesday, October 4, 2011 - 4:06:15 PM
Last modification on : Tuesday, April 2, 2019 - 1:42:41 AM

File

DASF_SIMBAD11.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-00628991, version 1

Citation

Emilie Morvant, Amaury Habrard, Stéphane Ayache. On the Usefulness of Similarity Based Projection Spaces for Transfer Learning. First International Workshop on Similarity-Based Pattern Recognition, Sep 2011, Venise, Italy. pp.1-16. ⟨hal-00628991⟩

Share

Metrics

Record views

462

Files downloads

324