Self-Taught Support Vector Machine

by   Parvin Razzaghi, et al.

In this paper, a new approach for classification of target task using limited labeled target data as well as enormous unlabeled source data is proposed which is called self-taught learning. The target and source data can be drawn from different distributions. In the previous approaches, covariate shift assumption is considered where the marginal distributions p(x) change over domains and the conditional distributions p(y|x) remain the same. In our approach, we propose a new objective function which simultaneously learns a common space T(.) where the conditional distributions over domains p(T(x)|y) remain the same and learns robust SVM classifiers for target task using both source and target data in the new representation. Hence, in the proposed objective function, the hidden label of the source data is also incorporated. We applied the proposed approach on Caltech-256, MSRC+LMO datasets and compared the performance of our algorithm to the available competing methods. Our method has a superior performance to the successful existing algorithms.


page 17

page 18

page 21

page 23

page 24

page 25

page 27


Transfer Learning with Label Noise

Transfer learning aims to improve learning in the target domain with lim...

Coupled Support Vector Machines for Supervised Domain Adaptation

Popular domain adaptation (DA) techniques learn a classifier for the tar...

Kernel Robust Bias-Aware Prediction under Covariate Shift

Under covariate shift, training (source) data and testing (target) data ...

Self-training Avoids Using Spurious Features Under Domain Shift

In unsupervised domain adaptation, existing theory focuses on situations...

Adversarial Domain Adaptation with Self-Training for EEG-based Sleep Stage Classification

Sleep staging is of great importance in the diagnosis and treatment of s...

Robust Fairness under Covariate Shift

Making predictions that are fair with regard to protected group membersh...

ℓ_p Slack Norm Support Vector Data Description

The support vector data description (SVDD) approach serves as a de facto...