On Optimal Learning With Random Features

Jiamin Liu; Heng Lian

doi:10.1109/TNNLS.2022.3152270

On Optimal Learning With Random Features

Jiamin Liu, Heng Lian^*

^*Corresponding author for this work

Research output: Journal Publications and Reviews › RGC 21 - Publication in refereed journal › peer-review

4 Citations (Scopus)

Abstract

We consider supervised learning in a reproducing kernel Hilbert space (RKHS) using random features. We show that the optimal rate is obtained under suitable regularity conditions, and at the same time improving on the existing bounds on the number of random features required. As a straightforward extension, distributed learning in the simple setting of one-shot communication is also considered that achieves the same optimal rate.

Original language	English
Pages (from-to)	9536-9541
Journal	IEEE Transactions on Neural Networks and Learning Systems
Volume	34
Issue number	11
Online published	2 Mar 2022
DOIs	https://doi.org/10.1109/TNNLS.2022.3152270
Publication status	Published - Nov 2023

Funding

The work of Heng Lian was supported in part by the NSFC and the Shenzhen Research Institute, City University of Hong Kong, under Project 11871411; and in part by the Hong Kong Research Grants Council (RGC) General Research Fund under Grant 11301718, Grant 11300519, Grant 11300721, and Grant 11311822.

Research Keywords

Convergence
Distributed learning
Hilbert space
Kernel
kernel method
optimal rate
random features
Standards
Supervised learning
Time complexity
Urban areas

Access Output

10.1109/TNNLS.2022.3152270

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

@article{c41ff7377ad54556921aa9b05f7be392,

title = "On Optimal Learning With Random Features",

abstract = "We consider supervised learning in a reproducing kernel Hilbert space (RKHS) using random features. We show that the optimal rate is obtained under suitable regularity conditions, and at the same time improving on the existing bounds on the number of random features required. As a straightforward extension, distributed learning in the simple setting of one-shot communication is also considered that achieves the same optimal rate.",

keywords = "Convergence, Distributed learning, Hilbert space, Kernel, kernel method, optimal rate, random features, Standards, Supervised learning, Time complexity, Urban areas",

author = "Jiamin Liu and Heng Lian",

year = "2023",

month = nov,

doi = "10.1109/TNNLS.2022.3152270",

language = "English",

volume = "34",

pages = "9536--9541",

journal = "IEEE Transactions on Neural Networks and Learning Systems",

issn = "2162-237X",

publisher = "IEEE Computational Intelligence Society",

number = "11",

}

TY - JOUR

T1 - On Optimal Learning With Random Features

AU - Liu, Jiamin

AU - Lian, Heng

PY - 2023/11

Y1 - 2023/11

N2 - We consider supervised learning in a reproducing kernel Hilbert space (RKHS) using random features. We show that the optimal rate is obtained under suitable regularity conditions, and at the same time improving on the existing bounds on the number of random features required. As a straightforward extension, distributed learning in the simple setting of one-shot communication is also considered that achieves the same optimal rate.

AB - We consider supervised learning in a reproducing kernel Hilbert space (RKHS) using random features. We show that the optimal rate is obtained under suitable regularity conditions, and at the same time improving on the existing bounds on the number of random features required. As a straightforward extension, distributed learning in the simple setting of one-shot communication is also considered that achieves the same optimal rate.

KW - Convergence

KW - Distributed learning

KW - Hilbert space

KW - Kernel

KW - kernel method

KW - optimal rate

KW - random features

KW - Standards

KW - Supervised learning

KW - Time complexity

KW - Urban areas

UR - http://www.scopus.com/inward/record.url?scp=85125720457&partnerID=8YFLogxK

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-85125720457&origin=recordpage

U2 - 10.1109/TNNLS.2022.3152270

DO - 10.1109/TNNLS.2022.3152270

M3 - RGC 21 - Publication in refereed journal

SN - 2162-237X

VL - 34

SP - 9536

EP - 9541

JO - IEEE Transactions on Neural Networks and Learning Systems

JF - IEEE Transactions on Neural Networks and Learning Systems

IS - 11

ER -

On Optimal Learning With Random Features

Abstract

Funding

Research Keywords

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

GRF: Low-rank Nonparametric Regression and Application to Reinforcement Learning

GRF: Distributed Estimation with Random Projection in Reproducing Kernel Hilbert Spaces

GRF: Low-rank tensor as a Dimension Reduction Tool in Complex Data Analysis

Cite this

On Optimal Learning With Random Features

Abstract

Funding

Research Keywords

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Projects

GRF: Low-rank Nonparametric Regression and Application to Reinforcement Learning

GRF: Distributed Estimation with Random Projection in Reproducing Kernel Hilbert Spaces

GRF: Low-rank tensor as a Dimension Reduction Tool in Complex Data Analysis

Cite this