Learning Curves and Bootstrap Estimates for Inference with Gaussian Processes: A Statistical Mechanics Study

Dorthe Malzahn; Manfred Opper

doi:10.1002/cplx.10082

Learning Curves and Bootstrap Estimates for Inference with Gaussian Processes: A Statistical Mechanics Study

Dorthe Malzahn
, Manfred Opper

Research output: Contribution to journal › Journal article › Research › peer-review

54 Downloads (Orbit)

Abstract

We use the replica method of statistical physics to study the average case performance of learning systems. The new feature of our theory is that general distributions of data can be treated, which enables applications to real data. For a class of Bayesian prediction models which are based on Gaussian processes, we discuss Bootstrap estimates for learning curves.

Original language	English
Journal	Complexity
Volume	8
Issue number	4
Pages (from-to)	57-63
ISSN	1076-2787
DOIs	https://doi.org/10.1002/cplx.10082
Publication status	Published - 2003

Access to Document

10.1002/cplx.10082

FulltextSubmitted manuscript, 182 KB

OpenUrl availability

Check availability in DTU Findit

Cite this

@article{ec5f49f831914c5e9ad9473b41647951,

title = "Learning Curves and Bootstrap Estimates for Inference with Gaussian Processes: A Statistical Mechanics Study",

abstract = "We use the replica method of statistical physics to study the average case performance of learning systems. The new feature of our theory is that general distributions of data can be treated, which enables applications to real data. For a class of Bayesian prediction models which are based on Gaussian processes, we discuss Bootstrap estimates for learning curves.",

keywords = "variational methods, Bootstrap, learning curves, statistical physics, Gaussian processes",

author = "Dorthe Malzahn and Manfred Opper",

year = "2003",

doi = "10.1002/cplx.10082",

language = "English",

volume = "8",

pages = "57--63",

journal = "Complexity",

issn = "1076-2787",

publisher = "John Wiley and Sons Inc.",

number = "4",

}

TY - JOUR

T1 - Learning Curves and Bootstrap Estimates for Inference with Gaussian Processes: A Statistical Mechanics Study

AU - Malzahn, Dorthe

AU - Opper, Manfred

PY - 2003

Y1 - 2003

N2 - We use the replica method of statistical physics to study the average case performance of learning systems. The new feature of our theory is that general distributions of data can be treated, which enables applications to real data. For a class of Bayesian prediction models which are based on Gaussian processes, we discuss Bootstrap estimates for learning curves.

AB - We use the replica method of statistical physics to study the average case performance of learning systems. The new feature of our theory is that general distributions of data can be treated, which enables applications to real data. For a class of Bayesian prediction models which are based on Gaussian processes, we discuss Bootstrap estimates for learning curves.

KW - variational methods

KW - Bootstrap

KW - learning curves

KW - statistical physics

KW - Gaussian processes

U2 - 10.1002/cplx.10082

DO - 10.1002/cplx.10082

M3 - Journal article

SN - 1076-2787

VL - 8

SP - 57

EP - 63

JO - Complexity

JF - Complexity

IS - 4

ER -

Learning Curves and Bootstrap Estimates for Inference with Gaussian Processes: A Statistical Mechanics Study

Abstract

Access to Document

OpenUrl availability

Fingerprint

Cite this