Collaborative multi-view denoising

Lei Zhang, Shupeng Wang, Xiaoyu Zhang, Yong Wang, Binbin Li, Dinggang Shen, Shuiwang Ji

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

7 Citations (Scopus)

Abstract

In multi-view learning applications, like multimedia analysis and information retrieval, we often encounter the corrupted view problem in which the data are corrupted by two different types of noises, i.e., the intra- and inter-view noises. The noises may affect these applications that commonly acquire complementary representations from different views. Therefore, how to denoise corrupted views from multi-view data is of great importance for applications that integrate and analyze representations from different views. However, the heterogeneity among multi-view representations brings a significant challenge on denoising corrupted views. To address this challenge, we propose a general framework to jointly denoise corrupted views in this paper. Specifically, aiming at capturing the semantic complementarity and distributional similarity among different views, a novel Heterogeneous Linear Metric Learning (HLML) model with low-rank regularization, leave-one-out validation, and pseudo-metric constraints is proposed. Our method linearly maps multiview data to a high-dimensional feature-homogeneous space that embeds the complementary information from different views. Furthermore, to remove the intra- and inter-view noises, we present a newMulti-view Semi-supervised Collaborative Denoising (MSCD) method with elementary transformation constraints and gradient energy competition to establish the complementary relationship among the heterogeneous representations. Experimental results demonstrate that our proposed methods are effective and efficient.

Original languageEnglish
Title of host publicationKDD 2016 - Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining
PublisherAssociation for Computing Machinery
Pages2045-2054
Number of pages10
Volume13-17-August-2016
ISBN (Electronic)9781450342322
DOIs
Publication statusPublished - 2016 Aug 13
Externally publishedYes
Event22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2016 - San Francisco, United States
Duration: 2016 Aug 132016 Aug 17

Other

Other22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2016
CountryUnited States
CitySan Francisco
Period16/8/1316/8/17

Fingerprint

Information retrieval
Semantics

Keywords

  • Denoising
  • Heterogeneity
  • Metric learning
  • Multi-view

ASJC Scopus subject areas

  • Software
  • Information Systems

Cite this

Zhang, L., Wang, S., Zhang, X., Wang, Y., Li, B., Shen, D., & Ji, S. (2016). Collaborative multi-view denoising. In KDD 2016 - Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (Vol. 13-17-August-2016, pp. 2045-2054). Association for Computing Machinery. https://doi.org/10.1145/2939672.2939811

Collaborative multi-view denoising. / Zhang, Lei; Wang, Shupeng; Zhang, Xiaoyu; Wang, Yong; Li, Binbin; Shen, Dinggang; Ji, Shuiwang.

KDD 2016 - Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Vol. 13-17-August-2016 Association for Computing Machinery, 2016. p. 2045-2054.

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

Zhang, L, Wang, S, Zhang, X, Wang, Y, Li, B, Shen, D & Ji, S 2016, Collaborative multi-view denoising. in KDD 2016 - Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. vol. 13-17-August-2016, Association for Computing Machinery, pp. 2045-2054, 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2016, San Francisco, United States, 16/8/13. https://doi.org/10.1145/2939672.2939811
Zhang L, Wang S, Zhang X, Wang Y, Li B, Shen D et al. Collaborative multi-view denoising. In KDD 2016 - Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Vol. 13-17-August-2016. Association for Computing Machinery. 2016. p. 2045-2054 https://doi.org/10.1145/2939672.2939811
Zhang, Lei ; Wang, Shupeng ; Zhang, Xiaoyu ; Wang, Yong ; Li, Binbin ; Shen, Dinggang ; Ji, Shuiwang. / Collaborative multi-view denoising. KDD 2016 - Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Vol. 13-17-August-2016 Association for Computing Machinery, 2016. pp. 2045-2054
@inproceedings{b6f119e4683d42338d322dfe69ca5509,
title = "Collaborative multi-view denoising",
abstract = "In multi-view learning applications, like multimedia analysis and information retrieval, we often encounter the corrupted view problem in which the data are corrupted by two different types of noises, i.e., the intra- and inter-view noises. The noises may affect these applications that commonly acquire complementary representations from different views. Therefore, how to denoise corrupted views from multi-view data is of great importance for applications that integrate and analyze representations from different views. However, the heterogeneity among multi-view representations brings a significant challenge on denoising corrupted views. To address this challenge, we propose a general framework to jointly denoise corrupted views in this paper. Specifically, aiming at capturing the semantic complementarity and distributional similarity among different views, a novel Heterogeneous Linear Metric Learning (HLML) model with low-rank regularization, leave-one-out validation, and pseudo-metric constraints is proposed. Our method linearly maps multiview data to a high-dimensional feature-homogeneous space that embeds the complementary information from different views. Furthermore, to remove the intra- and inter-view noises, we present a newMulti-view Semi-supervised Collaborative Denoising (MSCD) method with elementary transformation constraints and gradient energy competition to establish the complementary relationship among the heterogeneous representations. Experimental results demonstrate that our proposed methods are effective and efficient.",
keywords = "Denoising, Heterogeneity, Metric learning, Multi-view",
author = "Lei Zhang and Shupeng Wang and Xiaoyu Zhang and Yong Wang and Binbin Li and Dinggang Shen and Shuiwang Ji",
year = "2016",
month = "8",
day = "13",
doi = "10.1145/2939672.2939811",
language = "English",
volume = "13-17-August-2016",
pages = "2045--2054",
booktitle = "KDD 2016 - Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining",
publisher = "Association for Computing Machinery",

}

TY - GEN

T1 - Collaborative multi-view denoising

AU - Zhang, Lei

AU - Wang, Shupeng

AU - Zhang, Xiaoyu

AU - Wang, Yong

AU - Li, Binbin

AU - Shen, Dinggang

AU - Ji, Shuiwang

PY - 2016/8/13

Y1 - 2016/8/13

N2 - In multi-view learning applications, like multimedia analysis and information retrieval, we often encounter the corrupted view problem in which the data are corrupted by two different types of noises, i.e., the intra- and inter-view noises. The noises may affect these applications that commonly acquire complementary representations from different views. Therefore, how to denoise corrupted views from multi-view data is of great importance for applications that integrate and analyze representations from different views. However, the heterogeneity among multi-view representations brings a significant challenge on denoising corrupted views. To address this challenge, we propose a general framework to jointly denoise corrupted views in this paper. Specifically, aiming at capturing the semantic complementarity and distributional similarity among different views, a novel Heterogeneous Linear Metric Learning (HLML) model with low-rank regularization, leave-one-out validation, and pseudo-metric constraints is proposed. Our method linearly maps multiview data to a high-dimensional feature-homogeneous space that embeds the complementary information from different views. Furthermore, to remove the intra- and inter-view noises, we present a newMulti-view Semi-supervised Collaborative Denoising (MSCD) method with elementary transformation constraints and gradient energy competition to establish the complementary relationship among the heterogeneous representations. Experimental results demonstrate that our proposed methods are effective and efficient.

AB - In multi-view learning applications, like multimedia analysis and information retrieval, we often encounter the corrupted view problem in which the data are corrupted by two different types of noises, i.e., the intra- and inter-view noises. The noises may affect these applications that commonly acquire complementary representations from different views. Therefore, how to denoise corrupted views from multi-view data is of great importance for applications that integrate and analyze representations from different views. However, the heterogeneity among multi-view representations brings a significant challenge on denoising corrupted views. To address this challenge, we propose a general framework to jointly denoise corrupted views in this paper. Specifically, aiming at capturing the semantic complementarity and distributional similarity among different views, a novel Heterogeneous Linear Metric Learning (HLML) model with low-rank regularization, leave-one-out validation, and pseudo-metric constraints is proposed. Our method linearly maps multiview data to a high-dimensional feature-homogeneous space that embeds the complementary information from different views. Furthermore, to remove the intra- and inter-view noises, we present a newMulti-view Semi-supervised Collaborative Denoising (MSCD) method with elementary transformation constraints and gradient energy competition to establish the complementary relationship among the heterogeneous representations. Experimental results demonstrate that our proposed methods are effective and efficient.

KW - Denoising

KW - Heterogeneity

KW - Metric learning

KW - Multi-view

UR - http://www.scopus.com/inward/record.url?scp=84985032691&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84985032691&partnerID=8YFLogxK

U2 - 10.1145/2939672.2939811

DO - 10.1145/2939672.2939811

M3 - Conference contribution

AN - SCOPUS:84985032691

VL - 13-17-August-2016

SP - 2045

EP - 2054

BT - KDD 2016 - Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining

PB - Association for Computing Machinery

ER -