Meta-learning

Research output: Chapter in Book/Report/Conference proceedingChapterAcademicpeer-review

Abstract

Meta-learning, or learning to learn, is the science of systematically observing how different machine learning approaches perform on a wide range of learning tasks, and then learning from this experience, or meta-data, to learn new tasks much faster than otherwise possible. Not only does this dramatically speed up and improve the design of machine learning pipelines or neural architectures, it also allows us to replace hand-engineered algorithms with novel approaches learned in a data-driven way. In this chapter, we provide an overview of the state of the art in this fascinating and continuously evolving field.
LanguageEnglish
Title of host publicationAutomatic Machine Learning
Subtitle of host publicationMethods, Systems, Challenges
EditorsFrank Hutter, Lars Kotthoff, Joaquin Vanschoren
Place of PublicationCham
PublisherSpringer
Chapter2
Pages39-61
Number of pages27
ISBN (Electronic)978-3-030-05318-5
ISBN (Print)978-3-030-05317-8
DOIs
StatePublished - 2019

Fingerprint

Learning systems
Metadata
Pipelines

Keywords

  • meta-learning
  • Automatic Machine Learning

Cite this

Vanschoren, J. (2019). Meta-learning. In F. Hutter, L. Kotthoff, & J. Vanschoren (Eds.), Automatic Machine Learning: Methods, Systems, Challenges (pp. 39-61). Cham: Springer. DOI: 10.1007/978-3-030-05318-5_2
Vanschoren, J./ Meta-learning. Automatic Machine Learning: Methods, Systems, Challenges. editor / Frank Hutter ; Lars Kotthoff ; Joaquin Vanschoren. Cham : Springer, 2019. pp. 39-61
@inbook{880b7405f28f4477b28265dfa85cdd62,
title = "Meta-learning",
abstract = "Meta-learning, or learning to learn, is the science of systematically observing how different machine learning approaches perform on a wide range of learning tasks, and then learning from this experience, or meta-data, to learn new tasks much faster than otherwise possible. Not only does this dramatically speed up and improve the design of machine learning pipelines or neural architectures, it also allows us to replace hand-engineered algorithms with novel approaches learned in a data-driven way. In this chapter, we provide an overview of the state of the art in this fascinating and continuously evolving field.",
keywords = "meta-learning, Automatic Machine Learning",
author = "J. Vanschoren",
year = "2019",
doi = "10.1007/978-3-030-05318-5_2",
language = "English",
isbn = "978-3-030-05317-8",
pages = "39--61",
editor = "Hutter, {Frank } and Lars Kotthoff and Joaquin Vanschoren",
booktitle = "Automatic Machine Learning",
publisher = "Springer",
address = "Germany",

}

Vanschoren, J 2019, Meta-learning. in F Hutter, L Kotthoff & J Vanschoren (eds), Automatic Machine Learning: Methods, Systems, Challenges. Springer, Cham, pp. 39-61. DOI: 10.1007/978-3-030-05318-5_2

Meta-learning. / Vanschoren, J.

Automatic Machine Learning: Methods, Systems, Challenges. ed. / Frank Hutter; Lars Kotthoff; Joaquin Vanschoren. Cham : Springer, 2019. p. 39-61.

Research output: Chapter in Book/Report/Conference proceedingChapterAcademicpeer-review

TY - CHAP

T1 - Meta-learning

AU - Vanschoren,J.

PY - 2019

Y1 - 2019

N2 - Meta-learning, or learning to learn, is the science of systematically observing how different machine learning approaches perform on a wide range of learning tasks, and then learning from this experience, or meta-data, to learn new tasks much faster than otherwise possible. Not only does this dramatically speed up and improve the design of machine learning pipelines or neural architectures, it also allows us to replace hand-engineered algorithms with novel approaches learned in a data-driven way. In this chapter, we provide an overview of the state of the art in this fascinating and continuously evolving field.

AB - Meta-learning, or learning to learn, is the science of systematically observing how different machine learning approaches perform on a wide range of learning tasks, and then learning from this experience, or meta-data, to learn new tasks much faster than otherwise possible. Not only does this dramatically speed up and improve the design of machine learning pipelines or neural architectures, it also allows us to replace hand-engineered algorithms with novel approaches learned in a data-driven way. In this chapter, we provide an overview of the state of the art in this fascinating and continuously evolving field.

KW - meta-learning

KW - Automatic Machine Learning

U2 - 10.1007/978-3-030-05318-5_2

DO - 10.1007/978-3-030-05318-5_2

M3 - Chapter

SN - 978-3-030-05317-8

SP - 39

EP - 61

BT - Automatic Machine Learning

PB - Springer

CY - Cham

ER -

Vanschoren J. Meta-learning. In Hutter F, Kotthoff L, Vanschoren J, editors, Automatic Machine Learning: Methods, Systems, Challenges. Cham: Springer. 2019. p. 39-61. Available from, DOI: 10.1007/978-3-030-05318-5_2