Exclusive Lasso for Multi-task Feature Selection

Yang Zhou, Rong Jin, Steven Chu–Hong Hoi
Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics, PMLR 9:988-995, 2010.

Abstract

We propose a novel group regularization which we call exclusive lasso. Unlike the group lasso regularizer that assumes co-varying variables in groups, the proposed exclusive lasso regularizer models the scenario when variables in the same group compete with each other. Analysis is presented to illustrate the properties of the proposed regularizer. We present a framework of kernel-based multi-task feature selection algorithm based on the proposed exclusive lasso regularizer. An efficient algorithm is derived to solve the related optimization problem. Experiments with document categorization show that our approach outperforms state-of-the-art algorithms for multi-task feature selection.

Cite this Paper

BibTeX
@InProceedings{pmlr-v9-zhou10a, title = {Exclusive Lasso for Multi-task Feature Selection}, author = {Zhou, Yang and Jin, Rong and Hoi, Steven Chu–Hong}, booktitle = {Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics}, pages = {988--995}, year = {2010}, editor = {Teh, Yee Whye and Titterington, Mike}, volume = {9}, series = {Proceedings of Machine Learning Research}, address = {Chia Laguna Resort, Sardinia, Italy}, month = {13--15 May}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v9/zhou10a/zhou10a.pdf}, url = {https://proceedings.mlr.press/v9/zhou10a.html}, abstract = {We propose a novel group regularization which we call exclusive lasso. Unlike the group lasso regularizer that assumes co-varying variables in groups, the proposed exclusive lasso regularizer models the scenario when variables in the same group compete with each other. Analysis is presented to illustrate the properties of the proposed regularizer. We present a framework of kernel-based multi-task feature selection algorithm based on the proposed exclusive lasso regularizer. An efficient algorithm is derived to solve the related optimization problem. Experiments with document categorization show that our approach outperforms state-of-the-art algorithms for multi-task feature selection.} }
Endnote
%0 Conference Paper %T Exclusive Lasso for Multi-task Feature Selection %A Yang Zhou %A Rong Jin %A Steven Chu–Hong Hoi %B Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2010 %E Yee Whye Teh %E Mike Titterington %F pmlr-v9-zhou10a %I PMLR %P 988--995 %U https://proceedings.mlr.press/v9/zhou10a.html %V 9 %X We propose a novel group regularization which we call exclusive lasso. Unlike the group lasso regularizer that assumes co-varying variables in groups, the proposed exclusive lasso regularizer models the scenario when variables in the same group compete with each other. Analysis is presented to illustrate the properties of the proposed regularizer. We present a framework of kernel-based multi-task feature selection algorithm based on the proposed exclusive lasso regularizer. An efficient algorithm is derived to solve the related optimization problem. Experiments with document categorization show that our approach outperforms state-of-the-art algorithms for multi-task feature selection.
RIS
TY - CPAPER TI - Exclusive Lasso for Multi-task Feature Selection AU - Yang Zhou AU - Rong Jin AU - Steven Chu–Hong Hoi BT - Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics DA - 2010/03/31 ED - Yee Whye Teh ED - Mike Titterington ID - pmlr-v9-zhou10a PB - PMLR DP - Proceedings of Machine Learning Research VL - 9 SP - 988 EP - 995 L1 - http://proceedings.mlr.press/v9/zhou10a/zhou10a.pdf UR - https://proceedings.mlr.press/v9/zhou10a.html AB - We propose a novel group regularization which we call exclusive lasso. Unlike the group lasso regularizer that assumes co-varying variables in groups, the proposed exclusive lasso regularizer models the scenario when variables in the same group compete with each other. Analysis is presented to illustrate the properties of the proposed regularizer. We present a framework of kernel-based multi-task feature selection algorithm based on the proposed exclusive lasso regularizer. An efficient algorithm is derived to solve the related optimization problem. Experiments with document categorization show that our approach outperforms state-of-the-art algorithms for multi-task feature selection. ER -
APA
Zhou, Y., Jin, R. & Hoi, S.C.. (2010). Exclusive Lasso for Multi-task Feature Selection. Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 9:988-995 Available from https://proceedings.mlr.press/v9/zhou10a.html.

Related Material