Feature selection for transfer learning

Authors:
Selen Uguroglu;Jaime Carbonell
Affiliations:
Language Technologies Institute, Carnegie Mellon University;Language Technologies Institute, Carnegie Mellon University
Venue:
ECML PKDD'11 Proceedings of the 2011 European conference on Machine learning and knowledge discovery in databases - Volume Part III
Year:
2011

Citing 10
Cited 1

Semidefinite programming

SIAM Review
Kernel independent component analysis

The Journal of Machine Learning Research
Integrating structured biological data by Kernel Maximum Mean Discrepancy

Bioinformatics
Estimating Location Using Wi-Fi

IEEE Intelligent Systems
Feature Selection for Local Learning Based Clustering

PAKDD '09 Proceedings of the 13th Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining
Domain adaptation with structural correspondence learning

EMNLP '06 Proceedings of the 2006 Conference on Empirical Methods in Natural Language Processing
Transfer learning via dimensionality reduction

AAAI'08 Proceedings of the 23rd national conference on Artificial intelligence - Volume 2
Domain adaptation via transfer component analysis

IJCAI'09 Proceedings of the 21st international jont conference on Artifical intelligence
Discriminative semi-supervised feature selection via manifold regularization

IJCAI'09 Proceedings of the 21st international jont conference on Artifical intelligence
A Survey on Transfer Learning

IEEE Transactions on Knowledge and Data Engineering

Triplex transfer learning: exploiting both shared and distinct concepts for text classification

Proceedings of the sixth ACM international conference on Web search and data mining

Quantified Score

Hi-index	0.00

Visualization

Abstract

Common assumption in most machine learning algorithms is that, labeled (source) data and unlabeled (target) data are sampled from the same distribution. However, many real world tasks violate this assumption: in temporal domains, feature distributions may vary over time, clinical studies may have sampling bias, or sometimes sufficient labeled data for the domain of interest does not exist, and labeled data from a related domain must be utilized. In such settings, knowing in which dimensions source and target data vary is extremely important to reduce the distance between domains and accurately transfer knowledge. In this paper, we present a novel method to identify variant and invariant features between two datasets. Our contribution is two fold: First, we present a novel transfer learning approach for domain adaptation, and second, we formalize the problem of finding differently distributed features as a convex optimization problem. Experimental studies on synthetic and benchmark real world datasets show that our approach outperform other transfer learning approaches, and it aids the prediction accuracy significantly.