Random feature-based online multi-kernel learning in environments with unknown dynamics

Yanning Shen, Tianyi Chen, Georgios B Giannakis

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Kernel-based methods exhibit well-documented performance in various nonlinear learning tasks. Most of them rely on a preselected kernel, whose prudent choice presumes task-specific prior information. Especially when the latter is not available, multi-kernel learning has gained popularity thanks to its flexibility in choosing kernels from a prescribed kernel dictionary. Leveraging the random feature approximation and its recent orthogonality-promoting variant, the present contribution develops a scalable multi-kernel learning scheme (termed Raker) to obtain the sought nonlinear learning function 'on the fly,' first for static environments. To further boost performance in dynamic environments, an adaptive multi-kernel learning scheme (termed AdaRaker) is developed. AdaRaker accounts not only for data-driven learning of kernel combination, but also for the unknown dynamics. Performance is analyzed in terms of both static and dynamic regrets. AdaRaker is uniquely capable of tracking nonlinear learning functions in environments with unknown dynamics, and with with analytic performance guarantees. Tests with synthetic and real datasets are carried out to showcase the effectiveness of the novel algorithms.

Original languageEnglish (US)
JournalJournal of Machine Learning Research
Volume20
StatePublished - Feb 1 2019

Fingerprint

kernel
Unknown
Glossaries
Performance Guarantee
Regret
Learning
Prior Information
Dynamic Environment
Orthogonality
Data-driven
Flexibility
Approximation

Keywords

  • Dynamic and adversarial environments
  • Multi-kernel learning
  • Online learning
  • Random features
  • Reproducing kernel Hilbert space

Cite this

Random feature-based online multi-kernel learning in environments with unknown dynamics. / Shen, Yanning; Chen, Tianyi; Giannakis, Georgios B.

In: Journal of Machine Learning Research, Vol. 20, 01.02.2019.

Research output: Contribution to journalArticle

@article{14e43bf6b8964203aea1f27f60bca658,
title = "Random feature-based online multi-kernel learning in environments with unknown dynamics",
abstract = "Kernel-based methods exhibit well-documented performance in various nonlinear learning tasks. Most of them rely on a preselected kernel, whose prudent choice presumes task-specific prior information. Especially when the latter is not available, multi-kernel learning has gained popularity thanks to its flexibility in choosing kernels from a prescribed kernel dictionary. Leveraging the random feature approximation and its recent orthogonality-promoting variant, the present contribution develops a scalable multi-kernel learning scheme (termed Raker) to obtain the sought nonlinear learning function 'on the fly,' first for static environments. To further boost performance in dynamic environments, an adaptive multi-kernel learning scheme (termed AdaRaker) is developed. AdaRaker accounts not only for data-driven learning of kernel combination, but also for the unknown dynamics. Performance is analyzed in terms of both static and dynamic regrets. AdaRaker is uniquely capable of tracking nonlinear learning functions in environments with unknown dynamics, and with with analytic performance guarantees. Tests with synthetic and real datasets are carried out to showcase the effectiveness of the novel algorithms.",
keywords = "Dynamic and adversarial environments, Multi-kernel learning, Online learning, Random features, Reproducing kernel Hilbert space",
author = "Yanning Shen and Tianyi Chen and Giannakis, {Georgios B}",
year = "2019",
month = "2",
day = "1",
language = "English (US)",
volume = "20",
journal = "Journal of Machine Learning Research",
issn = "1532-4435",
publisher = "Microtome Publishing",

}

TY - JOUR

T1 - Random feature-based online multi-kernel learning in environments with unknown dynamics

AU - Shen, Yanning

AU - Chen, Tianyi

AU - Giannakis, Georgios B

PY - 2019/2/1

Y1 - 2019/2/1

N2 - Kernel-based methods exhibit well-documented performance in various nonlinear learning tasks. Most of them rely on a preselected kernel, whose prudent choice presumes task-specific prior information. Especially when the latter is not available, multi-kernel learning has gained popularity thanks to its flexibility in choosing kernels from a prescribed kernel dictionary. Leveraging the random feature approximation and its recent orthogonality-promoting variant, the present contribution develops a scalable multi-kernel learning scheme (termed Raker) to obtain the sought nonlinear learning function 'on the fly,' first for static environments. To further boost performance in dynamic environments, an adaptive multi-kernel learning scheme (termed AdaRaker) is developed. AdaRaker accounts not only for data-driven learning of kernel combination, but also for the unknown dynamics. Performance is analyzed in terms of both static and dynamic regrets. AdaRaker is uniquely capable of tracking nonlinear learning functions in environments with unknown dynamics, and with with analytic performance guarantees. Tests with synthetic and real datasets are carried out to showcase the effectiveness of the novel algorithms.

AB - Kernel-based methods exhibit well-documented performance in various nonlinear learning tasks. Most of them rely on a preselected kernel, whose prudent choice presumes task-specific prior information. Especially when the latter is not available, multi-kernel learning has gained popularity thanks to its flexibility in choosing kernels from a prescribed kernel dictionary. Leveraging the random feature approximation and its recent orthogonality-promoting variant, the present contribution develops a scalable multi-kernel learning scheme (termed Raker) to obtain the sought nonlinear learning function 'on the fly,' first for static environments. To further boost performance in dynamic environments, an adaptive multi-kernel learning scheme (termed AdaRaker) is developed. AdaRaker accounts not only for data-driven learning of kernel combination, but also for the unknown dynamics. Performance is analyzed in terms of both static and dynamic regrets. AdaRaker is uniquely capable of tracking nonlinear learning functions in environments with unknown dynamics, and with with analytic performance guarantees. Tests with synthetic and real datasets are carried out to showcase the effectiveness of the novel algorithms.

KW - Dynamic and adversarial environments

KW - Multi-kernel learning

KW - Online learning

KW - Random features

KW - Reproducing kernel Hilbert space

UR - http://www.scopus.com/inward/record.url?scp=85069459112&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85069459112&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:85069459112

VL - 20

JO - Journal of Machine Learning Research

JF - Journal of Machine Learning Research

SN - 1532-4435

ER -