Kernel extrapolation

S. V.N. Vishwanathan; Karsten M. Borgwardt; Omri Guttman; Alex Smola

doi:10.1016/j.neucom.2005.12.113

Kernel extrapolation

S. V.N. Vishwanathan^*, Karsten M. Borgwardt, Omri Guttman, Alex Smola

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

9 Citations (Scopus)

Abstract

We present a framework for efficient extrapolation of reduced rank approximations, graph kernels, and locally linear embeddings (LLE) to unseen data. We also present a principled method to combine many of these kernels and then extrapolate them. Central to our method is a theorem for matrix approximation, and an extension of the representer theorem to handle multiple joint regularization constraints. Experiments in protein classification demonstrate the feasibility of our approach.

Original language	English
Pages (from-to)	721-729
Number of pages	9
Journal	Neurocomputing
Volume	69
Issue number	7-9 SPEC. ISS.
DOIs	https://doi.org/10.1016/j.neucom.2005.12.113
Publication status	Published - Mar 2006

Access to Document

10.1016/j.neucom.2005.12.113

Cite this

@article{442564b477324c8c8ce7c1b4e16d8593,

title = "Kernel extrapolation",

abstract = "We present a framework for efficient extrapolation of reduced rank approximations, graph kernels, and locally linear embeddings (LLE) to unseen data. We also present a principled method to combine many of these kernels and then extrapolate them. Central to our method is a theorem for matrix approximation, and an extension of the representer theorem to handle multiple joint regularization constraints. Experiments in protein classification demonstrate the feasibility of our approach.",

keywords = "Graph kernels, Kernel methods, Protein classification, Regularization",

author = "Vishwanathan, {S. V.N.} and Borgwardt, {Karsten M.} and Omri Guttman and Alex Smola",

year = "2006",

month = mar,

doi = "10.1016/j.neucom.2005.12.113",

language = "English",

volume = "69",

pages = "721--729",

journal = "Neurocomputing",

issn = "0925-2312",

publisher = "Elsevier B.V.",

number = "7-9 SPEC. ISS.",

}

TY - JOUR

T1 - Kernel extrapolation

AU - Vishwanathan, S. V.N.

AU - Borgwardt, Karsten M.

AU - Guttman, Omri

AU - Smola, Alex

PY - 2006/3

Y1 - 2006/3

N2 - We present a framework for efficient extrapolation of reduced rank approximations, graph kernels, and locally linear embeddings (LLE) to unseen data. We also present a principled method to combine many of these kernels and then extrapolate them. Central to our method is a theorem for matrix approximation, and an extension of the representer theorem to handle multiple joint regularization constraints. Experiments in protein classification demonstrate the feasibility of our approach.

AB - We present a framework for efficient extrapolation of reduced rank approximations, graph kernels, and locally linear embeddings (LLE) to unseen data. We also present a principled method to combine many of these kernels and then extrapolate them. Central to our method is a theorem for matrix approximation, and an extension of the representer theorem to handle multiple joint regularization constraints. Experiments in protein classification demonstrate the feasibility of our approach.

KW - Graph kernels

KW - Kernel methods

KW - Protein classification

KW - Regularization

UR - http://www.scopus.com/inward/record.url?scp=32544442498&partnerID=8YFLogxK

U2 - 10.1016/j.neucom.2005.12.113

DO - 10.1016/j.neucom.2005.12.113

M3 - Article

SN - 0925-2312

VL - 69

SP - 721

EP - 729

JO - Neurocomputing

JF - Neurocomputing

IS - 7-9 SPEC. ISS.

ER -

Kernel extrapolation

Abstract

Access to Document

Other files and links

Fingerprint

Cite this