Online minimum error entropy algorithm with unbounded sampling

Cheng Wang; Ting Hu

doi:10.1142/S0219530518500148

Online minimum error entropy algorithm with unbounded sampling

Cheng Wang
, Ting Hu

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

13 Scopus citations

Abstract

Minimum error entropy (MEE) criterion is an important optimization method in information theoretic learning (ITL) and has been widely used and studied in various practical scenarios. In this paper, we shall introduce the online MEE algorithm for dealing with big datasets, associated with reproducing kernel Hilbert spaces (RKHS) and unbounded sampling processes. Explicit convergence rate will be given under the conditions of regularity of the regression function and polynomially decaying step sizes. Besides its low complexity, we will also show that the learning ability of online MEE is superior to the previous work in the literature. Our main techniques depend on integral operators on RKHS and probability inequalities for random variables with values in a Hilbert space.

Original language	English
Pages (from-to)	293-322
Number of pages	30
Journal	Analysis and Applications
Volume	17
Issue number	2
DOIs	https://doi.org/10.1142/S0219530518500148
State	Published - 1 Mar 2019
Externally published	Yes

Keywords

Information theoretic learning
minimum error entropy
online learning
pairwise learning
reproducing kernel Hilbert spaces

Access to Document

10.1142/S0219530518500148

Cite this

@article{1f66c58d270d4f09984d900b9e0bf981,

title = "Online minimum error entropy algorithm with unbounded sampling",

abstract = "Minimum error entropy (MEE) criterion is an important optimization method in information theoretic learning (ITL) and has been widely used and studied in various practical scenarios. In this paper, we shall introduce the online MEE algorithm for dealing with big datasets, associated with reproducing kernel Hilbert spaces (RKHS) and unbounded sampling processes. Explicit convergence rate will be given under the conditions of regularity of the regression function and polynomially decaying step sizes. Besides its low complexity, we will also show that the learning ability of online MEE is superior to the previous work in the literature. Our main techniques depend on integral operators on RKHS and probability inequalities for random variables with values in a Hilbert space.",

keywords = "Information theoretic learning, minimum error entropy, online learning, pairwise learning, reproducing kernel Hilbert spaces",

author = "Cheng Wang and Ting Hu",

note = "Publisher Copyright: {\textcopyright} 2019 World Scientific Publishing Company.",

year = "2019",

month = mar,

day = "1",

doi = "10.1142/S0219530518500148",

language = "英语",

volume = "17",

pages = "293--322",

journal = "Analysis and Applications",

issn = "0219-5305",

publisher = "World Scientific",

number = "2",

}

TY - JOUR

T1 - Online minimum error entropy algorithm with unbounded sampling

AU - Wang, Cheng

AU - Hu, Ting

PY - 2019/3/1

Y1 - 2019/3/1

N2 - Minimum error entropy (MEE) criterion is an important optimization method in information theoretic learning (ITL) and has been widely used and studied in various practical scenarios. In this paper, we shall introduce the online MEE algorithm for dealing with big datasets, associated with reproducing kernel Hilbert spaces (RKHS) and unbounded sampling processes. Explicit convergence rate will be given under the conditions of regularity of the regression function and polynomially decaying step sizes. Besides its low complexity, we will also show that the learning ability of online MEE is superior to the previous work in the literature. Our main techniques depend on integral operators on RKHS and probability inequalities for random variables with values in a Hilbert space.

AB - Minimum error entropy (MEE) criterion is an important optimization method in information theoretic learning (ITL) and has been widely used and studied in various practical scenarios. In this paper, we shall introduce the online MEE algorithm for dealing with big datasets, associated with reproducing kernel Hilbert spaces (RKHS) and unbounded sampling processes. Explicit convergence rate will be given under the conditions of regularity of the regression function and polynomially decaying step sizes. Besides its low complexity, we will also show that the learning ability of online MEE is superior to the previous work in the literature. Our main techniques depend on integral operators on RKHS and probability inequalities for random variables with values in a Hilbert space.

KW - Information theoretic learning

KW - minimum error entropy

KW - online learning

KW - pairwise learning

KW - reproducing kernel Hilbert spaces

UR - https://www.scopus.com/pages/publications/85062875568

U2 - 10.1142/S0219530518500148

DO - 10.1142/S0219530518500148

M3 - 文章

AN - SCOPUS:85062875568

SN - 0219-5305

VL - 17

SP - 293

EP - 322

JO - Analysis and Applications

JF - Analysis and Applications

IS - 2

ER -

Online minimum error entropy algorithm with unbounded sampling

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this