Mathematical Analysis on Information-Theoretic Metric Learning with Application to Supervised Learning

Jooyeon Choi; Chohong Min; Byungjoon Lee

doi:10.1109/ACCESS.2019.2937973

Mathematical Analysis on Information-Theoretic Metric Learning with Application to Supervised Learning

Jooyeon Choi, Chohong Min, Byungjoon Lee

Department of Mathematics

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

1 Scopus citations

Abstract

This article presents a concrete mathematical analysis on Information-Theoretic Metric Learning (ITML). The analysis provides a theoretical foundation for ITML, by supplying well-posedness, strong duality, and convergence. Our analysis suggests the correction of a typo in the original ITML article that may lead to the loss of accuracy in the metric learning. The necessity of this correction is confirmed by several numerical experiments on supervised learning.

Original language	English
Article number	8818103
Pages (from-to)	121998-122005
Number of pages	8
Journal	IEEE Access
Volume	7
DOIs	https://doi.org/10.1109/ACCESS.2019.2937973
State	Published - 2019

Bibliographical note

Publisher Copyright:
© 2013 IEEE.

Keywords

Bregman iteration
convex optimization
machine learning algorithm
mathematical analysis
metric learning

Access to Document

10.1109/ACCESS.2019.2937973

Cite this

@article{1561f6a6931f46d7a5348a58358a33ea,

title = "Mathematical Analysis on Information-Theoretic Metric Learning with Application to Supervised Learning",

abstract = "This article presents a concrete mathematical analysis on Information-Theoretic Metric Learning (ITML). The analysis provides a theoretical foundation for ITML, by supplying well-posedness, strong duality, and convergence. Our analysis suggests the correction of a typo in the original ITML article that may lead to the loss of accuracy in the metric learning. The necessity of this correction is confirmed by several numerical experiments on supervised learning.",

keywords = "Bregman iteration, convex optimization, machine learning algorithm, mathematical analysis, metric learning",

author = "Jooyeon Choi and Chohong Min and Byungjoon Lee",

note = "Publisher Copyright: {\textcopyright} 2013 IEEE.",

year = "2019",

doi = "10.1109/ACCESS.2019.2937973",

language = "English",

volume = "7",

pages = "121998--122005",

journal = "IEEE Access",

issn = "2169-3536",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - JOUR

T1 - Mathematical Analysis on Information-Theoretic Metric Learning with Application to Supervised Learning

AU - Choi, Jooyeon

AU - Min, Chohong

AU - Lee, Byungjoon

PY - 2019

Y1 - 2019

N2 - This article presents a concrete mathematical analysis on Information-Theoretic Metric Learning (ITML). The analysis provides a theoretical foundation for ITML, by supplying well-posedness, strong duality, and convergence. Our analysis suggests the correction of a typo in the original ITML article that may lead to the loss of accuracy in the metric learning. The necessity of this correction is confirmed by several numerical experiments on supervised learning.

AB - This article presents a concrete mathematical analysis on Information-Theoretic Metric Learning (ITML). The analysis provides a theoretical foundation for ITML, by supplying well-posedness, strong duality, and convergence. Our analysis suggests the correction of a typo in the original ITML article that may lead to the loss of accuracy in the metric learning. The necessity of this correction is confirmed by several numerical experiments on supervised learning.

KW - Bregman iteration

KW - convex optimization

KW - machine learning algorithm

KW - mathematical analysis

KW - metric learning

UR - http://www.scopus.com/inward/record.url?scp=85097341139&partnerID=8YFLogxK

U2 - 10.1109/ACCESS.2019.2937973

DO - 10.1109/ACCESS.2019.2937973

M3 - Article

AN - SCOPUS:85097341139

SN - 2169-3536

VL - 7

SP - 121998

EP - 122005

JO - IEEE Access

JF - IEEE Access

M1 - 8818103

ER -

Mathematical Analysis on Information-Theoretic Metric Learning with Application to Supervised Learning

Abstract

Bibliographical note

Keywords

Access to Document

Fingerprint

Cite this