Label Information Guided Graph Construction for Semi-Supervised Learning

Liansheng Zhuang, Zihan Zhou, Shenghua Gao, Jingwen Yin, Zhouchen Lin, Yi Ma

Research output: Contribution to journalArticle

20 Citations (Scopus)

Abstract

In the literature, most existing graph-based semi-supervised learning methods only use the label information of observed samples in the label propagation stage, while ignoring such valuable information when learning the graph. In this paper, we argue that it is beneficial to consider the label information in the graph learning stage. Specifically, by enforcing the weight of edges between labeled samples of different classes to be zero, we explicitly incorporate the label information into the state-of-the-art graph learning methods, such as the low-rank representation (LRR), and propose a novel semi-supervised graph learning method called semi-supervised low-rank representation. This results in a convex optimization problem with linear constraints, which can be solved by the linearized alternating direction method. Though we take LRR as an example, our proposed method is in fact very general and can be applied to any self-representation graph learning methods. Experiment results on both synthetic and real data sets demonstrate that the proposed graph learning method can better capture the global geometric structure of the data, and therefore is more effective for semi-supervised learning tasks.

Original languageEnglish (US)
Article number7931596
Pages (from-to)4182-4192
Number of pages11
JournalIEEE Transactions on Image Processing
Volume26
Issue number9
DOIs
StatePublished - Sep 2017

Fingerprint

Supervised learning
Labels
Convex optimization
Experiments

All Science Journal Classification (ASJC) codes

  • Software
  • Computer Graphics and Computer-Aided Design

Cite this

Zhuang, Liansheng ; Zhou, Zihan ; Gao, Shenghua ; Yin, Jingwen ; Lin, Zhouchen ; Ma, Yi. / Label Information Guided Graph Construction for Semi-Supervised Learning. In: IEEE Transactions on Image Processing. 2017 ; Vol. 26, No. 9. pp. 4182-4192.
@article{b1fac00b9486413d9796df60511f9e61,
title = "Label Information Guided Graph Construction for Semi-Supervised Learning",
abstract = "In the literature, most existing graph-based semi-supervised learning methods only use the label information of observed samples in the label propagation stage, while ignoring such valuable information when learning the graph. In this paper, we argue that it is beneficial to consider the label information in the graph learning stage. Specifically, by enforcing the weight of edges between labeled samples of different classes to be zero, we explicitly incorporate the label information into the state-of-the-art graph learning methods, such as the low-rank representation (LRR), and propose a novel semi-supervised graph learning method called semi-supervised low-rank representation. This results in a convex optimization problem with linear constraints, which can be solved by the linearized alternating direction method. Though we take LRR as an example, our proposed method is in fact very general and can be applied to any self-representation graph learning methods. Experiment results on both synthetic and real data sets demonstrate that the proposed graph learning method can better capture the global geometric structure of the data, and therefore is more effective for semi-supervised learning tasks.",
author = "Liansheng Zhuang and Zihan Zhou and Shenghua Gao and Jingwen Yin and Zhouchen Lin and Yi Ma",
year = "2017",
month = "9",
doi = "10.1109/TIP.2017.2703120",
language = "English (US)",
volume = "26",
pages = "4182--4192",
journal = "IEEE Transactions on Image Processing",
issn = "1057-7149",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "9",

}

Label Information Guided Graph Construction for Semi-Supervised Learning. / Zhuang, Liansheng; Zhou, Zihan; Gao, Shenghua; Yin, Jingwen; Lin, Zhouchen; Ma, Yi.

In: IEEE Transactions on Image Processing, Vol. 26, No. 9, 7931596, 09.2017, p. 4182-4192.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Label Information Guided Graph Construction for Semi-Supervised Learning

AU - Zhuang, Liansheng

AU - Zhou, Zihan

AU - Gao, Shenghua

AU - Yin, Jingwen

AU - Lin, Zhouchen

AU - Ma, Yi

PY - 2017/9

Y1 - 2017/9

N2 - In the literature, most existing graph-based semi-supervised learning methods only use the label information of observed samples in the label propagation stage, while ignoring such valuable information when learning the graph. In this paper, we argue that it is beneficial to consider the label information in the graph learning stage. Specifically, by enforcing the weight of edges between labeled samples of different classes to be zero, we explicitly incorporate the label information into the state-of-the-art graph learning methods, such as the low-rank representation (LRR), and propose a novel semi-supervised graph learning method called semi-supervised low-rank representation. This results in a convex optimization problem with linear constraints, which can be solved by the linearized alternating direction method. Though we take LRR as an example, our proposed method is in fact very general and can be applied to any self-representation graph learning methods. Experiment results on both synthetic and real data sets demonstrate that the proposed graph learning method can better capture the global geometric structure of the data, and therefore is more effective for semi-supervised learning tasks.

AB - In the literature, most existing graph-based semi-supervised learning methods only use the label information of observed samples in the label propagation stage, while ignoring such valuable information when learning the graph. In this paper, we argue that it is beneficial to consider the label information in the graph learning stage. Specifically, by enforcing the weight of edges between labeled samples of different classes to be zero, we explicitly incorporate the label information into the state-of-the-art graph learning methods, such as the low-rank representation (LRR), and propose a novel semi-supervised graph learning method called semi-supervised low-rank representation. This results in a convex optimization problem with linear constraints, which can be solved by the linearized alternating direction method. Though we take LRR as an example, our proposed method is in fact very general and can be applied to any self-representation graph learning methods. Experiment results on both synthetic and real data sets demonstrate that the proposed graph learning method can better capture the global geometric structure of the data, and therefore is more effective for semi-supervised learning tasks.

UR - http://www.scopus.com/inward/record.url?scp=85028436469&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85028436469&partnerID=8YFLogxK

U2 - 10.1109/TIP.2017.2703120

DO - 10.1109/TIP.2017.2703120

M3 - Article

AN - SCOPUS:85028436469

VL - 26

SP - 4182

EP - 4192

JO - IEEE Transactions on Image Processing

JF - IEEE Transactions on Image Processing

SN - 1057-7149

IS - 9

M1 - 7931596

ER -