Multi-Instance Kernels

T Gartner; PA Flach; A Kowalczyk; AJ Smola

Multi-Instance Kernels

T Gartner, PA Flach, A Kowalczyk, AJ Smola

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

Abstract

Learning from structured data is becoming increasingly important. However, most prior work on kernel methods has focused on learning from attribute-value data. Only recently, research started investigating kernels for structured data. This paper considers kernels for multi-instance problems - a class of concepts on individuals represented by sets. The main result of this paper is a kernel on multi-instance data that can be shown to separate positive and negative sets under natural assumptions. This kernel compares favorably with state of the art multi-instance learning algorithms in an empirical study. Finally, we give some concluding remarks and propose future work that might further improve the results.

Translated title of the contribution	Multi-Instance Kernels
Original language	English
Title of host publication	Unknown
Editors	Claude Sammut, Achim Hoffmann
Publisher	Morgan Kaufmann
Pages	179 - 186
Number of pages	7
ISBN (Print)	1558608737
Publication status	Published - Jul 2002

Bibliographical note

Conference Proceedings/Title of Journal: Proceedings of the 19th International Conference on Machine Learning

Access to Document

http://www.cs.bris.ac.uk/Publications/pub_info.jsp?id=1000688

Handle.net

Persistent link

Cite this

@inproceedings{2a5630a3e38242d1a3bf9736e35ddc81,

title = "Multi-Instance Kernels",

abstract = "Learning from structured data is becoming increasingly important. However, most prior work on kernel methods has focused on learning from attribute-value data. Only recently, research started investigating kernels for structured data. This paper considers kernels for multi-instance problems - a class of concepts on individuals represented by sets. The main result of this paper is a kernel on multi-instance data that can be shown to separate positive and negative sets under natural assumptions. This kernel compares favorably with state of the art multi-instance learning algorithms in an empirical study. Finally, we give some concluding remarks and propose future work that might further improve the results. ",

author = "T Gartner and PA Flach and A Kowalczyk and AJ Smola",

note = "Conference Proceedings/Title of Journal: Proceedings of the 19th International Conference on Machine Learning",

year = "2002",

month = jul,

language = "English",

isbn = "1558608737",

pages = "179 -- 186",

editor = "Claude Sammut and Achim Hoffmann",

booktitle = "Unknown",

publisher = "Morgan Kaufmann",

}

TY - GEN

T1 - Multi-Instance Kernels

AU - Gartner, T

AU - Flach, PA

AU - Kowalczyk, A

AU - Smola, AJ

N1 - Conference Proceedings/Title of Journal: Proceedings of the 19th International Conference on Machine Learning

PY - 2002/7

Y1 - 2002/7

N2 - Learning from structured data is becoming increasingly important. However, most prior work on kernel methods has focused on learning from attribute-value data. Only recently, research started investigating kernels for structured data. This paper considers kernels for multi-instance problems - a class of concepts on individuals represented by sets. The main result of this paper is a kernel on multi-instance data that can be shown to separate positive and negative sets under natural assumptions. This kernel compares favorably with state of the art multi-instance learning algorithms in an empirical study. Finally, we give some concluding remarks and propose future work that might further improve the results.

AB - Learning from structured data is becoming increasingly important. However, most prior work on kernel methods has focused on learning from attribute-value data. Only recently, research started investigating kernels for structured data. This paper considers kernels for multi-instance problems - a class of concepts on individuals represented by sets. The main result of this paper is a kernel on multi-instance data that can be shown to separate positive and negative sets under natural assumptions. This kernel compares favorably with state of the art multi-instance learning algorithms in an empirical study. Finally, we give some concluding remarks and propose future work that might further improve the results.

M3 - Conference Contribution (Conference Proceeding)

SN - 1558608737

SP - 179

EP - 186

BT - Unknown

A2 - Sammut, Claude

A2 - Hoffmann, Achim

PB - Morgan Kaufmann

ER -

Multi-Instance Kernels

Abstract

Bibliographical note

Access to Document

Handle.net

Fingerprint

Cite this