Connection between SVM+ and multi-task learning

Lichen Liang; Vladimir S Cherkassky

doi:10.1109/IJCNN.2008.4634079

Connection between SVM+ and multi-task learning

Lichen Liang, Vladimir S Cherkassky

Electrical and Computer Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

43 Scopus citations

Abstract

Exploiting additional information to improve traditional inductive learning is an active research in machine learning. When data are naturally separated into groups, SVM+[7] can effectively utilize this structure information to improve generalization. Alternatively, we can view learning based on data from each group as an individual task, but all these tasks are somehow related; so the same problem can also be formulated as a multi-task learning problem. Following the SVM+ approach, we propose a new multi-task learning algorithm called svm+MTL, which can be thought as an adaptation of SVM+ for solving MTL problem. The connections between SVM+ and svm+MTL are discussed and their performance is compared using synthetic data sets.

Original language	English (US)
Title of host publication	2008 International Joint Conference on Neural Networks, IJCNN 2008
Pages	2048-2054
Number of pages	7
DOIs	https://doi.org/10.1109/IJCNN.2008.4634079
State	Published - 2008
Event	2008 International Joint Conference on Neural Networks, IJCNN 2008 - Hong Kong, China Duration: Jun 1 2008 → Jun 8 2008

Publication series

Name	Proceedings of the International Joint Conference on Neural Networks

Other

Other	2008 International Joint Conference on Neural Networks, IJCNN 2008
Country/Territory	China
City	Hong Kong
Period	6/1/08 → 6/8/08

Publisher link

10.1109/IJCNN.2008.4634079

Find It

Find It at U of M Libraries

Cite this

Liang, L & Cherkassky, VS 2008, Connection between SVM+ and multi-task learning. in 2008 International Joint Conference on Neural Networks, IJCNN 2008., 4634079, Proceedings of the International Joint Conference on Neural Networks, pp. 2048-2054, 2008 International Joint Conference on Neural Networks, IJCNN 2008, Hong Kong, China, 6/1/08. https://doi.org/10.1109/IJCNN.2008.4634079

@inproceedings{010818a0be4548b987b092d99b1943ec,

title = "Connection between SVM+ and multi-task learning",

abstract = "Exploiting additional information to improve traditional inductive learning is an active research in machine learning. When data are naturally separated into groups, SVM+[7] can effectively utilize this structure information to improve generalization. Alternatively, we can view learning based on data from each group as an individual task, but all these tasks are somehow related; so the same problem can also be formulated as a multi-task learning problem. Following the SVM+ approach, we propose a new multi-task learning algorithm called svm+MTL, which can be thought as an adaptation of SVM+ for solving MTL problem. The connections between SVM+ and svm+MTL are discussed and their performance is compared using synthetic data sets.",

author = "Lichen Liang and Cherkassky, {Vladimir S}",

year = "2008",

doi = "10.1109/IJCNN.2008.4634079",

language = "English (US)",

isbn = "9781424418213",

series = "Proceedings of the International Joint Conference on Neural Networks",

pages = "2048--2054",

booktitle = "2008 International Joint Conference on Neural Networks, IJCNN 2008",

note = "2008 International Joint Conference on Neural Networks, IJCNN 2008 ; Conference date: 01-06-2008 Through 08-06-2008",

}

TY - GEN

T1 - Connection between SVM+ and multi-task learning

AU - Liang, Lichen

AU - Cherkassky, Vladimir S

PY - 2008

Y1 - 2008

N2 - Exploiting additional information to improve traditional inductive learning is an active research in machine learning. When data are naturally separated into groups, SVM+[7] can effectively utilize this structure information to improve generalization. Alternatively, we can view learning based on data from each group as an individual task, but all these tasks are somehow related; so the same problem can also be formulated as a multi-task learning problem. Following the SVM+ approach, we propose a new multi-task learning algorithm called svm+MTL, which can be thought as an adaptation of SVM+ for solving MTL problem. The connections between SVM+ and svm+MTL are discussed and their performance is compared using synthetic data sets.

AB - Exploiting additional information to improve traditional inductive learning is an active research in machine learning. When data are naturally separated into groups, SVM+[7] can effectively utilize this structure information to improve generalization. Alternatively, we can view learning based on data from each group as an individual task, but all these tasks are somehow related; so the same problem can also be formulated as a multi-task learning problem. Following the SVM+ approach, we propose a new multi-task learning algorithm called svm+MTL, which can be thought as an adaptation of SVM+ for solving MTL problem. The connections between SVM+ and svm+MTL are discussed and their performance is compared using synthetic data sets.

UR - http://www.scopus.com/inward/record.url?scp=56349168083&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=56349168083&partnerID=8YFLogxK

U2 - 10.1109/IJCNN.2008.4634079

DO - 10.1109/IJCNN.2008.4634079

M3 - Conference contribution

AN - SCOPUS:56349168083

SN - 9781424418213

T3 - Proceedings of the International Joint Conference on Neural Networks

SP - 2048

EP - 2054

BT - 2008 International Joint Conference on Neural Networks, IJCNN 2008

T2 - 2008 International Joint Conference on Neural Networks, IJCNN 2008

Y2 - 1 June 2008 through 8 June 2008

ER -

Connection between SVM+ and multi-task learning

Abstract

Publication series

Other

Publisher link

Find It

Other files and links

Fingerprint

Cite this