Kernel design using boosting

Koby Crammer; Joseph Keshet; Yoram Singer

Kernel design using boosting

Koby Crammer, Joseph Keshet, Yoram Singer

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

Abstract

The focus of the paper is the problem of learning kernel operators from empirical data. We cast the kernel design problem as the construction of an accurate kernel from simple (and less accurate) base kernels. We use the boosting paradigm to perform the kernel construction process. To do so, we modify the booster so as to accommodate kernel operators. We also devise an efficient weak-learner for simple kernels that is based on generalized eigen vector decomposition. We demonstrate the effectiveness of our approach on synthetic data and on the USPS dataset. On the USPS dataset, the performance of the Perceptron algorithm with learned kernels is systematically better than a fixed RBF kernel.

Original language	English (US)
Title of host publication	Advances in Neural Information Processing Systems 15 - Proceedings of the 2002 Conference, NIPS 2002
Publisher	Neural information processing systems foundation
ISBN (Print)	0262025507, 9780262025508
State	Published - 2003
Event	16th Annual Neural Information Processing Systems Conference, NIPS 2002 - Vancouver, BC, Canada Duration: Dec 9 2002 → Dec 14 2002

Publication series

Name	Advances in Neural Information Processing Systems
ISSN (Print)	1049-5258

Other

Other	16th Annual Neural Information Processing Systems Conference, NIPS 2002
Country/Territory	Canada
City	Vancouver, BC
Period	12/9/02 → 12/14/02

All Science Journal Classification (ASJC) codes

Computer Networks and Communications
Information Systems
Signal Processing

Cite this

@inproceedings{4067c12e838546078449fa6b2656f21b,

title = "Kernel design using boosting",

abstract = "The focus of the paper is the problem of learning kernel operators from empirical data. We cast the kernel design problem as the construction of an accurate kernel from simple (and less accurate) base kernels. We use the boosting paradigm to perform the kernel construction process. To do so, we modify the booster so as to accommodate kernel operators. We also devise an efficient weak-learner for simple kernels that is based on generalized eigen vector decomposition. We demonstrate the effectiveness of our approach on synthetic data and on the USPS dataset. On the USPS dataset, the performance of the Perceptron algorithm with learned kernels is systematically better than a fixed RBF kernel.",

author = "Koby Crammer and Joseph Keshet and Yoram Singer",

year = "2003",

language = "English (US)",

isbn = "0262025507",

series = "Advances in Neural Information Processing Systems",

publisher = "Neural information processing systems foundation",

booktitle = "Advances in Neural Information Processing Systems 15 - Proceedings of the 2002 Conference, NIPS 2002",

note = "16th Annual Neural Information Processing Systems Conference, NIPS 2002 ; Conference date: 09-12-2002 Through 14-12-2002",

}

Crammer, K, Keshet, J & Singer, Y 2003, Kernel design using boosting. in Advances in Neural Information Processing Systems 15 - Proceedings of the 2002 Conference, NIPS 2002. Advances in Neural Information Processing Systems, Neural information processing systems foundation, 16th Annual Neural Information Processing Systems Conference, NIPS 2002, Vancouver, BC, Canada, 12/9/02.

Kernel design using boosting. / Crammer, Koby; Keshet, Joseph; Singer, Yoram.
Advances in Neural Information Processing Systems 15 - Proceedings of the 2002 Conference, NIPS 2002. Neural information processing systems foundation, 2003. (Advances in Neural Information Processing Systems).

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

TY - GEN

T1 - Kernel design using boosting

AU - Crammer, Koby

AU - Keshet, Joseph

AU - Singer, Yoram

PY - 2003

Y1 - 2003

N2 - The focus of the paper is the problem of learning kernel operators from empirical data. We cast the kernel design problem as the construction of an accurate kernel from simple (and less accurate) base kernels. We use the boosting paradigm to perform the kernel construction process. To do so, we modify the booster so as to accommodate kernel operators. We also devise an efficient weak-learner for simple kernels that is based on generalized eigen vector decomposition. We demonstrate the effectiveness of our approach on synthetic data and on the USPS dataset. On the USPS dataset, the performance of the Perceptron algorithm with learned kernels is systematically better than a fixed RBF kernel.

AB - The focus of the paper is the problem of learning kernel operators from empirical data. We cast the kernel design problem as the construction of an accurate kernel from simple (and less accurate) base kernels. We use the boosting paradigm to perform the kernel construction process. To do so, we modify the booster so as to accommodate kernel operators. We also devise an efficient weak-learner for simple kernels that is based on generalized eigen vector decomposition. We demonstrate the effectiveness of our approach on synthetic data and on the USPS dataset. On the USPS dataset, the performance of the Perceptron algorithm with learned kernels is systematically better than a fixed RBF kernel.

UR - http://www.scopus.com/inward/record.url?scp=84898956003&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84898956003&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84898956003

SN - 0262025507

SN - 9780262025508

T3 - Advances in Neural Information Processing Systems

BT - Advances in Neural Information Processing Systems 15 - Proceedings of the 2002 Conference, NIPS 2002

PB - Neural information processing systems foundation

T2 - 16th Annual Neural Information Processing Systems Conference, NIPS 2002

Y2 - 9 December 2002 through 14 December 2002

ER -

Kernel design using boosting

Abstract

Publication series

Other

All Science Journal Classification (ASJC) codes

Other files and links

Fingerprint

Cite this