Learning manifold representation from multimodal data for event detection in flickr-like social media

Zhenguo Yang; Qing Li; Wenyin Liu; Yun Ma

doi:10.1007/978-3-319-32055-7_14

Learning manifold representation from multimodal data for event detection in flickr-like social media

Zhenguo Yang
, Qing Li
, Wenyin Liu
, Yun Ma

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

3 Citations (Scopus)

Abstract

In this work, a three-stage social event detection model is devised to discover events in Flickr data. As the features possessed by the data are typically heterogeneous, a multimodal fusion model (M²F) exploits a soft-voting strategy and a reinforcing model is devised to learn fused features in the first stage. Furthermore, a Laplacian non-negative matrix factorization (LNMF) model is exploited to extract compact manifold representation. Particularly, a Laplacian regularization term constructed on the multimodal features is introduced to keep the geometry structure of the data. Finally, clustering algorithms can be applied seamlessly in order to detect event clusters. Extensive experiments conducted on the real-world dataset reveal the M²F-LNMF-based approaches outperform the baselines.

Original language	English
Title of host publication	Database Systems for Advanced Applications - DASFAA 2016 International Workshops
Subtitle of host publication	BDMS, BDQM, MoI, and SeCoP, Proceedings
Editors	Jinho Kim, Hong Gao, Yasushi Sakurai
Publisher	Springer-Verlag
Pages	160-167
Number of pages	8
ISBN (Print)	9783319320540
DOIs	https://doi.org/10.1007/978-3-319-32055-7_14
Publication status	Published - 1 Jan 2016
Externally published	Yes
Event	International Workshop on Database Systems for Advanced Applications, DASFAA 2016 - Dallas, United States Duration: 16 Apr 2016 → 19 Apr 2016

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	9645
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	International Workshop on Database Systems for Advanced Applications, DASFAA 2016
Country/Territory	United States
City	Dallas
Period	16/04/16 → 19/04/16

Keywords

Event detection
Manifold learning
Multimedia content analysis
Multimodal fusion
Social media analytics

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

More information

10.1007/978-3-319-32055-7_14

Cite this

Yang, Z., Li, Q., Liu, W., & Ma, Y. (2016). Learning manifold representation from multimodal data for event detection in flickr-like social media. In J. Kim, H. Gao, & Y. Sakurai (Eds.), Database Systems for Advanced Applications - DASFAA 2016 International Workshops: BDMS, BDQM, MoI, and SeCoP, Proceedings (pp. 160-167). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9645). Springer-Verlag. https://doi.org/10.1007/978-3-319-32055-7_14

Yang, Zhenguo ; Li, Qing ; Liu, Wenyin et al. / Learning manifold representation from multimodal data for event detection in flickr-like social media. Database Systems for Advanced Applications - DASFAA 2016 International Workshops: BDMS, BDQM, MoI, and SeCoP, Proceedings. editor / Jinho Kim ; Hong Gao ; Yasushi Sakurai. Springer-Verlag, 2016. pp. 160-167 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{1bcd6407decd43d88d3531d3f4aeb89b,

title = "Learning manifold representation from multimodal data for event detection in flickr-like social media",

abstract = "In this work, a three-stage social event detection model is devised to discover events in Flickr data. As the features possessed by the data are typically heterogeneous, a multimodal fusion model (M2F) exploits a soft-voting strategy and a reinforcing model is devised to learn fused features in the first stage. Furthermore, a Laplacian non-negative matrix factorization (LNMF) model is exploited to extract compact manifold representation. Particularly, a Laplacian regularization term constructed on the multimodal features is introduced to keep the geometry structure of the data. Finally, clustering algorithms can be applied seamlessly in order to detect event clusters. Extensive experiments conducted on the real-world dataset reveal the M2F-LNMF-based approaches outperform the baselines.",

keywords = "Event detection, Manifold learning, Multimedia content analysis, Multimodal fusion, Social media analytics",

author = "Zhenguo Yang and Qing Li and Wenyin Liu and Yun Ma",

year = "2016",

month = jan,

day = "1",

doi = "10.1007/978-3-319-32055-7\_14",

language = "English",

isbn = "9783319320540",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer-Verlag",

pages = "160--167",

editor = "Jinho Kim and Hong Gao and Yasushi Sakurai",

booktitle = "Database Systems for Advanced Applications - DASFAA 2016 International Workshops",

note = "International Workshop on Database Systems for Advanced Applications, DASFAA 2016 ; Conference date: 16-04-2016 Through 19-04-2016",

}

Yang, Z, Li, Q, Liu, W & Ma, Y 2016, Learning manifold representation from multimodal data for event detection in flickr-like social media. in J Kim, H Gao & Y Sakurai (eds), Database Systems for Advanced Applications - DASFAA 2016 International Workshops: BDMS, BDQM, MoI, and SeCoP, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9645, Springer-Verlag, pp. 160-167, International Workshop on Database Systems for Advanced Applications, DASFAA 2016, Dallas, United States, 16/04/16. https://doi.org/10.1007/978-3-319-32055-7_14

Learning manifold representation from multimodal data for event detection in flickr-like social media. / Yang, Zhenguo; Li, Qing; Liu, Wenyin et al.
Database Systems for Advanced Applications - DASFAA 2016 International Workshops: BDMS, BDQM, MoI, and SeCoP, Proceedings. ed. / Jinho Kim; Hong Gao; Yasushi Sakurai. Springer-Verlag, 2016. p. 160-167 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9645).

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

TY - GEN

T1 - Learning manifold representation from multimodal data for event detection in flickr-like social media

AU - Yang, Zhenguo

AU - Li, Qing

AU - Liu, Wenyin

AU - Ma, Yun

PY - 2016/1/1

Y1 - 2016/1/1

N2 - In this work, a three-stage social event detection model is devised to discover events in Flickr data. As the features possessed by the data are typically heterogeneous, a multimodal fusion model (M2F) exploits a soft-voting strategy and a reinforcing model is devised to learn fused features in the first stage. Furthermore, a Laplacian non-negative matrix factorization (LNMF) model is exploited to extract compact manifold representation. Particularly, a Laplacian regularization term constructed on the multimodal features is introduced to keep the geometry structure of the data. Finally, clustering algorithms can be applied seamlessly in order to detect event clusters. Extensive experiments conducted on the real-world dataset reveal the M2F-LNMF-based approaches outperform the baselines.

AB - In this work, a three-stage social event detection model is devised to discover events in Flickr data. As the features possessed by the data are typically heterogeneous, a multimodal fusion model (M2F) exploits a soft-voting strategy and a reinforcing model is devised to learn fused features in the first stage. Furthermore, a Laplacian non-negative matrix factorization (LNMF) model is exploited to extract compact manifold representation. Particularly, a Laplacian regularization term constructed on the multimodal features is introduced to keep the geometry structure of the data. Finally, clustering algorithms can be applied seamlessly in order to detect event clusters. Extensive experiments conducted on the real-world dataset reveal the M2F-LNMF-based approaches outperform the baselines.

KW - Event detection

KW - Manifold learning

KW - Multimedia content analysis

KW - Multimodal fusion

KW - Social media analytics

UR - https://www.scopus.com/pages/publications/84964088560

U2 - 10.1007/978-3-319-32055-7_14

DO - 10.1007/978-3-319-32055-7_14

M3 - Conference article published in proceeding or book

AN - SCOPUS:84964088560

SN - 9783319320540

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 160

EP - 167

BT - Database Systems for Advanced Applications - DASFAA 2016 International Workshops

A2 - Kim, Jinho

A2 - Gao, Hong

A2 - Sakurai, Yasushi

PB - Springer-Verlag

T2 - International Workshop on Database Systems for Advanced Applications, DASFAA 2016

Y2 - 16 April 2016 through 19 April 2016

ER -

Yang Z, Li Q, Liu W, Ma Y. Learning manifold representation from multimodal data for event detection in flickr-like social media. In Kim J, Gao H, Sakurai Y, editors, Database Systems for Advanced Applications - DASFAA 2016 International Workshops: BDMS, BDQM, MoI, and SeCoP, Proceedings. Springer-Verlag. 2016. p. 160-167. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-32055-7_14

Learning manifold representation from multimodal data for event detection in flickr-like social media

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

More information

Other files and links

Fingerprint

Cite this