Distance metric learning with eigenvalue fine tuning

Wenquan Wang, Ya Zhang, Takayuki Furuzuki

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Distance metric learning focuses on learning one global or multiple local distance functions to draw similar instances close to each other and push away dissimilar ones. Most existing work has to do matrix projection to learn distance functions. In this paper, we present a novel distance function learning model which is based on eigenvalue fine tuning. Our model not only is able to learn the global distance function but also can be easily adopted into local metric learning tasks. From the perspective of dimension reduction, the proposed model can measure how much information has been preserved after feature transformation. Moreover, we connect our model with principal components analysis to improve its performance by introducing the label information. Experimental results have demonstrated the effectiveness of the proposed method.

Original languageEnglish
Title of host publication2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages502-509
Number of pages8
Volume2017-May
ISBN (Electronic)9781509061815
DOIs
Publication statusPublished - 2017 Jun 30
Event2017 International Joint Conference on Neural Networks, IJCNN 2017 - Anchorage, United States
Duration: 2017 May 142017 May 19

Other

Other2017 International Joint Conference on Neural Networks, IJCNN 2017
CountryUnited States
CityAnchorage
Period17/5/1417/5/19

Fingerprint

Tuning
Principal component analysis
Labels

ASJC Scopus subject areas

  • Software
  • Artificial Intelligence

Cite this

Wang, W., Zhang, Y., & Furuzuki, T. (2017). Distance metric learning with eigenvalue fine tuning. In 2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings (Vol. 2017-May, pp. 502-509). [7965895] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IJCNN.2017.7965895

Distance metric learning with eigenvalue fine tuning. / Wang, Wenquan; Zhang, Ya; Furuzuki, Takayuki.

2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings. Vol. 2017-May Institute of Electrical and Electronics Engineers Inc., 2017. p. 502-509 7965895.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Wang, W, Zhang, Y & Furuzuki, T 2017, Distance metric learning with eigenvalue fine tuning. in 2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings. vol. 2017-May, 7965895, Institute of Electrical and Electronics Engineers Inc., pp. 502-509, 2017 International Joint Conference on Neural Networks, IJCNN 2017, Anchorage, United States, 17/5/14. https://doi.org/10.1109/IJCNN.2017.7965895
Wang W, Zhang Y, Furuzuki T. Distance metric learning with eigenvalue fine tuning. In 2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings. Vol. 2017-May. Institute of Electrical and Electronics Engineers Inc. 2017. p. 502-509. 7965895 https://doi.org/10.1109/IJCNN.2017.7965895
Wang, Wenquan ; Zhang, Ya ; Furuzuki, Takayuki. / Distance metric learning with eigenvalue fine tuning. 2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings. Vol. 2017-May Institute of Electrical and Electronics Engineers Inc., 2017. pp. 502-509
@inproceedings{993289ec3cb144cca7e611d8896a7b34,
title = "Distance metric learning with eigenvalue fine tuning",
abstract = "Distance metric learning focuses on learning one global or multiple local distance functions to draw similar instances close to each other and push away dissimilar ones. Most existing work has to do matrix projection to learn distance functions. In this paper, we present a novel distance function learning model which is based on eigenvalue fine tuning. Our model not only is able to learn the global distance function but also can be easily adopted into local metric learning tasks. From the perspective of dimension reduction, the proposed model can measure how much information has been preserved after feature transformation. Moreover, we connect our model with principal components analysis to improve its performance by introducing the label information. Experimental results have demonstrated the effectiveness of the proposed method.",
author = "Wenquan Wang and Ya Zhang and Takayuki Furuzuki",
year = "2017",
month = "6",
day = "30",
doi = "10.1109/IJCNN.2017.7965895",
language = "English",
volume = "2017-May",
pages = "502--509",
booktitle = "2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
address = "United States",

}

TY - GEN

T1 - Distance metric learning with eigenvalue fine tuning

AU - Wang, Wenquan

AU - Zhang, Ya

AU - Furuzuki, Takayuki

PY - 2017/6/30

Y1 - 2017/6/30

N2 - Distance metric learning focuses on learning one global or multiple local distance functions to draw similar instances close to each other and push away dissimilar ones. Most existing work has to do matrix projection to learn distance functions. In this paper, we present a novel distance function learning model which is based on eigenvalue fine tuning. Our model not only is able to learn the global distance function but also can be easily adopted into local metric learning tasks. From the perspective of dimension reduction, the proposed model can measure how much information has been preserved after feature transformation. Moreover, we connect our model with principal components analysis to improve its performance by introducing the label information. Experimental results have demonstrated the effectiveness of the proposed method.

AB - Distance metric learning focuses on learning one global or multiple local distance functions to draw similar instances close to each other and push away dissimilar ones. Most existing work has to do matrix projection to learn distance functions. In this paper, we present a novel distance function learning model which is based on eigenvalue fine tuning. Our model not only is able to learn the global distance function but also can be easily adopted into local metric learning tasks. From the perspective of dimension reduction, the proposed model can measure how much information has been preserved after feature transformation. Moreover, we connect our model with principal components analysis to improve its performance by introducing the label information. Experimental results have demonstrated the effectiveness of the proposed method.

UR - http://www.scopus.com/inward/record.url?scp=85031022670&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85031022670&partnerID=8YFLogxK

U2 - 10.1109/IJCNN.2017.7965895

DO - 10.1109/IJCNN.2017.7965895

M3 - Conference contribution

AN - SCOPUS:85031022670

VL - 2017-May

SP - 502

EP - 509

BT - 2017 International Joint Conference on Neural Networks, IJCNN 2017 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

ER -