Distribution matching for transduction

Novi Quadrianto; James Petterson; Alex J. Smola

Distribution matching for transduction

Novi Quadrianto^*, James Petterson, Alex J. Smola

^*Corresponding author for this work

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

22 Citations (Scopus)

Abstract

Many transductive inference algorithms assume that distributions over training and test estimates should be related, e.g. by providing a large margin of separation on both sets. We use this idea to design a transduction algorithm which can be used without modification for classification, regression, and structured estimation. At its heart we exploit the fact that for a good learner the distributions over the outputs on training and test sets should match. This is a classical two-sample problem which can be solved efficiently in its most general form by using distance measures in Hilbert Space. It turns out that a number of existing heuristics can be viewed as special cases of our approach.

Original language	English
Title of host publication	Advances in Neural Information Processing Systems 22 - Proceedings of the 2009 Conference
Publisher	Neural Information Processing Systems
Pages	1500-1508
Number of pages	9
ISBN (Print)	9781615679119
Publication status	Published - 2009
Event	23rd Annual Conference on Neural Information Processing Systems, NIPS 2009 - Vancouver, BC, Canada Duration: 7 Dec 2009 → 10 Dec 2009

Publication series

Name	Advances in Neural Information Processing Systems 22 - Proceedings of the 2009 Conference

Conference

Conference	23rd Annual Conference on Neural Information Processing Systems, NIPS 2009
Country/Territory	Canada
City	Vancouver, BC
Period	7/12/09 → 10/12/09

Cite this

@inproceedings{f09ad950e4784649a15d749117f5bba2,

title = "Distribution matching for transduction",

abstract = "Many transductive inference algorithms assume that distributions over training and test estimates should be related, e.g. by providing a large margin of separation on both sets. We use this idea to design a transduction algorithm which can be used without modification for classification, regression, and structured estimation. At its heart we exploit the fact that for a good learner the distributions over the outputs on training and test sets should match. This is a classical two-sample problem which can be solved efficiently in its most general form by using distance measures in Hilbert Space. It turns out that a number of existing heuristics can be viewed as special cases of our approach.",

author = "Novi Quadrianto and James Petterson and Smola, {Alex J.}",

year = "2009",

language = "English",

isbn = "9781615679119",

series = "Advances in Neural Information Processing Systems 22 - Proceedings of the 2009 Conference",

publisher = "Neural Information Processing Systems",

pages = "1500--1508",

booktitle = "Advances in Neural Information Processing Systems 22 - Proceedings of the 2009 Conference",

note = "23rd Annual Conference on Neural Information Processing Systems, NIPS 2009 ; Conference date: 07-12-2009 Through 10-12-2009",

}

Quadrianto, N, Petterson, J & Smola, AJ 2009, Distribution matching for transduction. in Advances in Neural Information Processing Systems 22 - Proceedings of the 2009 Conference. Advances in Neural Information Processing Systems 22 - Proceedings of the 2009 Conference, Neural Information Processing Systems, pp. 1500-1508, 23rd Annual Conference on Neural Information Processing Systems, NIPS 2009, Vancouver, BC, Canada, 7/12/09.

Distribution matching for transduction. / Quadrianto, Novi; Petterson, James; Smola, Alex J.
Advances in Neural Information Processing Systems 22 - Proceedings of the 2009 Conference. Neural Information Processing Systems, 2009. p. 1500-1508 (Advances in Neural Information Processing Systems 22 - Proceedings of the 2009 Conference).

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

TY - GEN

T1 - Distribution matching for transduction

AU - Quadrianto, Novi

AU - Petterson, James

AU - Smola, Alex J.

PY - 2009

Y1 - 2009

N2 - Many transductive inference algorithms assume that distributions over training and test estimates should be related, e.g. by providing a large margin of separation on both sets. We use this idea to design a transduction algorithm which can be used without modification for classification, regression, and structured estimation. At its heart we exploit the fact that for a good learner the distributions over the outputs on training and test sets should match. This is a classical two-sample problem which can be solved efficiently in its most general form by using distance measures in Hilbert Space. It turns out that a number of existing heuristics can be viewed as special cases of our approach.

AB - Many transductive inference algorithms assume that distributions over training and test estimates should be related, e.g. by providing a large margin of separation on both sets. We use this idea to design a transduction algorithm which can be used without modification for classification, regression, and structured estimation. At its heart we exploit the fact that for a good learner the distributions over the outputs on training and test sets should match. This is a classical two-sample problem which can be solved efficiently in its most general form by using distance measures in Hilbert Space. It turns out that a number of existing heuristics can be viewed as special cases of our approach.

UR - http://www.scopus.com/inward/record.url?scp=77956528369&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9781615679119

T3 - Advances in Neural Information Processing Systems 22 - Proceedings of the 2009 Conference

SP - 1500

EP - 1508

BT - Advances in Neural Information Processing Systems 22 - Proceedings of the 2009 Conference

PB - Neural Information Processing Systems

T2 - 23rd Annual Conference on Neural Information Processing Systems, NIPS 2009

Y2 - 7 December 2009 through 10 December 2009

ER -

Distribution matching for transduction

Abstract

Publication series

Conference

Other files and links

Fingerprint

Cite this