Relational Data Community Discovery and Learning

Relational Data Community Discovery and Learning

Sponsor: National Science Foundation

This proposed project addresses a three year integrated research and education program focusing on engaging into an in-depth research on a series of fundamental, open, but very important issues leading to relational data community discovery and learning, built upon our existing strength on the state-of-the-art research on this topic.

The intellectual merit of this project includes the revolutionized understanding of the unsupervised general relational data clustering and learning as well as the expected breakthrough in the community discovery and learning methodologies that shall undoubtedly advance the literature of data mining and machine learning and generate profound impact in the related areas.

The broader impacts of this project are two folds. Educationally, the development, the implementation, and the evaluation of the innovative community outreach activities proposed in this project shall promote the timely and effective knowledge dissemination related to relational data mining and machine learning and shall further enrich the pedagogical literature; the disseminated knowledge to the collaborating parties, especially the collaborating high school, shall further advance and enhance the high school education services and syllabi and develop the model for high schools' research and services to the whole society. Technologically, the expected breakthrough in developing the novel theory on relational data community discovery and learning shall embrace a new era of technological revolution in a wide range of applications in the world, and in particular, shall benefit the collaborating organizations in developing and advancing their domain expertise in applications related to social network mining in general and Web data mining in particular.

It is well-observed that the whole world is full of data, and is also highly related in terms of the different types of the data objects such as people, organizations, and events. In many applications, it is intended to discover the hidden structures through such relationships involving different types of data objects in the world, in addition to "clusters" of the same type of data objects. For example, in financial services, it is often needed to identify any potential fraud activities reflected in the normal transactions that involve people and financial institutions; in commercial sales, it is often needed to link the customer purchase patterns to the potential sales promotion strategies to identify what kinds of customers are related to what kinds of commercial products through what kinds of service providers; in Web search industries, it is extremely desirable to identify what kinds of users use what kinds of Web pages and are highly influenced by what kinds of advertisements related to what kinds of commercial industries.

On the other hand, it is also true that it is too often that we do not have the luxury to have any training data with ground truth for knowledge discovery. Consequently, unsupervised relational data learning is expected and desired for all these situations.

In this research, we focus on the most general scenario of relational data: the data objects may have attributes, homogeneous relations (among data objects of the same type) and/or heterogeneous relations (between data objects of different types). Given such general relational data, all the practical situations are considered as the special cases of this general scenario, and thus the novel unified theory as well as the related methodologies we wish to develop in this research shall be applicable to any real-world relational data knowledge discovery problems, potentially resulting in revolutionary technology development and making the proposed work fundamentally new and uniquely distinct from all the existing literature. Consequently, we define a relational data community in the broad sense that includes not only the local clusters of the same type of data objects, but more importantly also the global, hidden structures incorporating relationships with different types of data objects.

Relational data community discovery and learning is a fairly new area with many challenging and fundamentally new issues completely open.

On the other hand, solutions to these issues may lead to revolutionary technology development that shall generate significant societal impacts.

The work to be accomplished in this project shall be radically new because it is based on innovative preliminary research and it is to address a set of fundamentally new problems with fundamentally new solutions that not only aim at developing a better in-depth understanding of the literature, but more importantly it is likely to generate revolutionary technology development with significant societal impacts. Specifically, this project focuses on the following three objectives to be achieved synergistically: (1) to address a series of challenging, fundamentally new, but very important issues on relational data community discovery and learning to lead to the development of a unified, fundamentally new theory on this topic to have a better in-depth understanding of the literature; (2) to extensively evaluate the theory and methodologies to be developed in collaborations with the domain experts in Web search industries as a specific application to social network mining; and (3) to develop and evaluate the innovative community outreach and education activities through the existing partnership with a local high school to further promote the knowledge dissemination from this research.

NSF Project Manager:  Dr. Maria Zemankova

Project Personnel:

PI: Prof. Zhongfei (Mark) Zhang

PhD students:

Pin Zhao
Zhen Guo
Yi Xu

NSF REU Students:

Greg Stoddard
Bryan Bernard
Philip Dexter
Kevin Hannon
Daniel F. McFadden

Partners:

Publications:

Ming Yang, Yingming Li, and Zhongfei (Mark) Zhang, Multi-Task Learning with Gaussian Matrix Generalized Inverse Gaussian Model, Proc. International Conference on Machine Learning, Atlanta, Georgia, USA, June, 2013, (20% acceptance rate)

Xinyan Lu, Fei Wu, Siliang Tang, Zhongfei (Mark) Zhang, Xiaofei He, and Yueting Zhuang, A Low Rank Structural Large Margin Method for Cross-Modal Ranking, Proc. The 36th Annual ACM SIGIR Conference, Dublin, Ireland, July, 2013, (19.9% acceptance rate)

Yingming Li, ZhongAng Qi, Zhongfei (Mark) Zhang, and Ming Yang, Learning with Limited and Noisy Tagging, Proc. The ACM International Conference on Multimedia, Barcelona, Catalunya, Spain, October, 2013, (20% acceptance rate)

Xinyan Lu, Fei Wu, Siliang Tang, Zhongfei (Mark) Zhang, Xiaofei He, and Yueting Zhuang, Cross-Media Semantic Representation via Bi-directional Learning to Rank, Proc. The ACM International Conference on Multimedia, Barcelona, Catalunya, Spain, October, 2013, (20% acceptance rate)

Xi Li, Anthony R. Dick, Chunhua Shen, Zhongfei Zhang, Anton van den Hengel, and Hanzi Wang, Visual Tracking With Spatio-Temporal Dempster-Shafer Information Fusion, IEEE Transactions on Image Processing, IEEE Signal Processing Society Press, Volune 22, Number 8, Pages 3028-3040, 2013

Zhen Guo, Zhongfei Zhang, Shenghuo Zhu, Yun Chi, Yihong Gong, A Two-Level Topic Model towards Knowledge Discovery from Citation Networks, IEEE Transactions on Knowledge and Data Engineering, IEEE Computer Society Press, accepted, 2013

Xi Li, Weiming Hu, Chunhua Shen, Zhongfei Zhang, Anthony R. Dick, and Anton van den Hengel, Context-Aware Hypergraph Construction for Robust Spectral Clustering, IEEE Transactions on Knowledge and Data Engineering, IEEE Computer Society Press, accepted, 2013

Zheng Fang and Zhongfei (Mark) Zhang, Discriminant Transfer Learning on Manifold, Proc. SIAM International Conference on Data Mining, Austin, TX, USA, May, 2013, (25.5% acceptance rate) [pdf]

Xi Li, Weiming Hu, Chunhua Shen, Zhongfei Zhang, Anthony Dick, and Anton van den Hengel, A Survey of Appearance Models in Visual Object Tracking, ACM Transactions on Intelligent Systems and Technology, ACM Press, accepted [pdf]

Zheng Fang and Zhongfei (Mark) Zhang, Simultaneously Combing Multi-View Multi-Label Learning with Maximum Margin Classification, Proc. IEEE International Conference on Data Mining, Brussels, Belgium, December, 2012, (9.3% acceptance rate)

Xi Li, Weiming Hu, Zhongfei Zhang, and David Suter, An Incremental DPMM-Based Method for Trajectory Clustering, Modeling and Retrieval, IEEE Transactions on Pattern Analysis and Machine Intelligence, IEEE Computer Society Press, accepted, 2012

Fei Wu, Yahong Han, Xiang Liu, Jian Shao, Yueting Zhuang, Zhongfei Zhang, The heterogeneous feature selection with structural sparsity for multimedia annotation and hashing: a survey, nternational Journal of Multimedia Information Retrieval, Springer, Volume 1, Number 1, Pages 3-15, April, 2012, DOI 10.1007/s13735-012-0001-9

Zhouzhou He, Zhongfei Zhang, Philip S. Yu, Overlapping Community Detection Combining Content and Link, Journal of Zhejiang University --- Part C: Computer Science, accepted, 2012

Weiming Hu, Haiqiang Zuo, Ou Wu, Yunfei Chen, Zhongfei Zhang, and David Suter, Single and Multiple Object Tracking Using Log-Euclidean Riemannian Subspace and Block-Division Appearance Model, IEEE Transactions on Pattern Analysis and Machine Intelligence, IEEE Computer Society Press, accepted, 2012

Tianbing Xu, Zhongfei Zhang, Philip S. Yu, and Bo Long, Generative Models for Evolutionary Clustering, ACM Transactions on Knowledge Discovery from Data, ACM Press, accepted, 2012

Zhongang Qi, Ming Yang, Zhongfei Zhang, Zhengyou Zhang, Multi-View Learning from Imperfect Tagging, Proc. the 20th ACM International Conference on Multimedia, Nara, Japan, 2012, (20% acceptance rate)

Zhongang Qi, Ming Yang, Zhongfei Zhang, Zhengyou Zhang, Mining Noisy Tagging from Multi-Label Space, Proc. the 21st ACM Conference on Information and Knowledge Management}, Maui, HW, USA, 2012, (14.3% acceptance rate)

Zhongfei (Mark) Zhang, Zhengyou Zhang, Ramesh Jain, and Yueting Zhuang, Guest Editors' Introduction: Special Section on Connected Multimedia, Journal of Multimedia, Academy Publisher, Volume 7, Number 1, 2012

Zhongfei Zhang, Zhengyou Zhang, Ramesh Jain, Yueting Zhuang, Noshir Contractor, Alexander G. Hauptmann, Alejandro (Alex) Jaimes, Wanqing Li, Alexander C. Loui, Tao Mei, Nicu Sebe, Yonghong Tian, Vincent S. Tseng, Qing Wang, Changsheng Xu, Huimin Yu, Shiwen Yu, Societally Connected Multimedia across Cultures, Journal of Zhejiang University --- Part C (Computer Science), Springer, accepted, 2012

Mingyu Fan, Xiaoqin Zhang;, Zhouchen Lin, Zhongfei Zhang and Hujun Bao, Geodesic based semi-supervised multi-manifold feature extraction, Proc. IEEE International Conference on Data Mining, Brussels, Belgium, December, 2012, (9.3% acceptance rate)

Jun Gao, Weiming Hu, Zhongfei (Mark) Zhang, Ou Wu, Unsupervised Ensemble Learning for Mining Top-n Outliers, Proc. the 16th Pacific-Asia Conference on Knowledge Discovery and Data Mining, Kuala Lumpur, Malaysia, May, 2012, (8.3% acceptance rate)

Yi Xu, Zhongfei (Mark) Zhang, Bo Long, and Philip S. Yu, Pattern Change Discovery between High Dimensional Datasets, Proc. the 20th ACM Conference on Information and Knowledge Management, Glasgow, UK, October, 2011, (15% acceptance rate)

Bo Long and Zhongfei (Mark) Zhang, A General Model for Relational Clustering, in Data Mining, Edited by Kimito Funatsu, IN-TECH, 2011

Zhongfei (Mark) Zhang and Ruofei Zhang, Multimedia Information Mining, in Multimedia Image and Video Processing, 2nd Edition, Edited by Ling Guan, Yifeng He, and Sun-Yuan Kung, CRC Press, 2011

Weiming Hu, Xi Li, Xiaoqin Zhang, Xinchu Shi, Stephen Maybank, and Zhongfei Zhang, Incremental Tensor Subspace Learning and Its Applications to Foreground Segmentation and Tracking, International Journal of Computer Vision, Springer, Volume 91, No. 3, pages 303-327, 2011

Zhongang Qi, Ming Yang, Zhongfei (Mark) Zhang, and Zhengyou Zhang, Mining Partially Annotated Images, Proc. the 17th ACM Conference on Knowledge Discovery and Data Mining, San Diego, CA, USA, August, 2011, (6.9% acceptance rate)

Jun Gao, Weiming Hu, Zhongfei (Mark) Zhang, Ou Wu, RKOF: Robust Kernel-Based Local Outlier Detection, Proc. the 15th Pacific-Asia Conference on Knowledge Discovery and Data Mining, Shenzhen, China, 2011, (17.5% acceptance rate)

Zhongfei Zhang and Haroon Khan, A Holistic, In-Compression Approach to Mining Independent Motion Segments for Massive Surveillance Video Collections, in Video Search and Mining, Edited by Dan Schonfeld, Caifeng Shan, Dacheng Tao, and Liang Wan, Springer, ISBN 978-3-642-12899-8, 2010

Zhongfei (Mark) Zhang and Ruofei Zhang, Multimedia Data Mining, in Data Mining and Knowledge Discovery Handbook, 2nd Ed., Edited by Oded Maimon and Lior Rokach, Springer, 2010

Zhongfei (Mark) Zhang, Bo Long, Zhen Guo, Tianbing Xu, and Philip S. Yu, Machine Learning Approaches to Link-Based Clustering, in Link Mining: Models, Algorithms and Applications, Edited by Philip S. Yu, Christos Faloutsos, and Jiawei Han, Springer, 2010

Zhen Guo, Zhongfei (Mark) Zhang, Eric P. Xing, and Christos Faloutsos, A Max Margin Framework on Image Annotation and Multimodal Image Retrieval, in Multimedia, Edited by Vedran Kordic, IN-TECH, 2010

Zhongfei (Mark) Zhang, Zhen Guo, Christos Faloutsos, Eric P. Xing, and Jia-Yu Pan, On the scalability and adaptability for multimodal image retrieval and image annotation, in Machine Learning Techniques for Adaptive Multimedia Retrieval: Technologies Applications and Perspectives, Edited by Roger Wei, Idea Group Inc., 2010

Xi Li, Weiming Hu, Hanzi Wang, and Zhongfei (Mark) Zhang, Robust Object Tracking Using A Spatial Pyramid Heat Kernel Structural Information Representation, Nurocomputing, Elsevier Science, Volume 73, No. 16-18, pages 3179-3190, October, 2010

Bo Long, Zhongfei (Mark) Zhang, and Philip S. Yu, A General Framework for Relation Graph Clustering, Knowledge and Information Systems, Springer, Volume 24, Number 3, Pages 393-413, September, 2010

Bo Long, Zhongfei (Mark) Zhang, and Philip S. Yu, Relational Data Clustering: Models, Algorithms, and Applications, Taylor & Francis/CRC Press, 2010, ISBN: 9781420072617

Zhongfei (Mark) Zhang, Zhen Guo, and Jia-Yu (Tim) Pan, A Multiple-Instance Learning Based Approach to Multimodal Data Mining, International Journal of Digital Library Systems, IGI Global, Volume 1, Number 2, Pages 23-41, April-June, 2010

Xi Li, Weiming Hu, Hanzi Wang, and Zhongfei (Mark) Zhang, Linear Discriminant Analysis Using Rotational Invariant L1 Norm, Nurocomputing, Elsevier Science, Volume 73, No. 13-15, pages 2571-2579, August, 2010

Xi Li, Weiming Hu, Zhongfei (Mark) Zhang, and Hanzi Wang, Heat Kernel Based Local Binary Pattern for Face Representation and Classification, IEEE Signal Processing Letters, IEEE Signal Processing Society Press, Volume 17, Issue 3, Pages 308-311, March, 2010

Jun Gao, Weiming Hu, Zhongfei (Mark) Zhang, Ou Wu, Local Outlier detection Based on Kernel Regression, Proc. International Conference on Pattern Recognition, Istanbul, Turkey, August, 2010, (54% acceptance rate)

Zhen Guo, Shenghuo Zhu, Yun Chi, Zhongfei (Mark) Zhang, and Yihong Gong, Unsupervised Learning from Linked Documents, Proc. International Conference on Pattern Recognition, Istanbul, Turkey, August, 2010, (54% acceptance rate)

Zhen Guo, Shenghuo Zhu, Zhongfei (Mark) Zhang, Yun Chi, and Yihong Gong, A Topic Model for Linked Documents and Update Rules for its Estimation, Proc. AAAI, Atlanta, GA, USA, July, 2010, (26.9% acceptance rate)

Zhen GuoBo Long, Zhongfei (Mark) Zhang, and Philip S. Yu, Relational Data Clustering: Models, Algorithms, and Applications, Taylor & Francis/CRC Press, 2009, ISBN: 9781420072617

Zhongfei Zhang and Haroon Khan, A Holistic, In-Compression Approach to Mining Independent Motion Segments for Massive Surveillance Video Collections, in Video Search and Mining, Edited by Dan Schonfeld, Caifeng Shan, Dacheng Tao, and Liang Wan, Springer, 2009

Zhongfei (Mark) Zhang and Ruofei Zhang, Multimedia Data Mining, in Data Mining and Knowledge Discovery Handbook, 2nd Ed., Edited by Oded Maimon and Lior Rokach, Springer, 2009

Zhongfei (Mark) Zhang, Bo Long, Zhen Guo, Tianbing Xu, and Philip S. Yu, Machine Learning Approaches to Link-Based Clustering, in Link Mining: Models, Algorithms and Applications, Edited by Philip S. Yu, Christos Faloutsos, and Jiawei Han, Springer, 2009

Zhen Guo, Zhongfei (Mark) Zhang, Eric P. Xing, and Christos Faloutsos, A Max Margin Framework on Image Annotation and Multimodal Image Retrieval, in Multimedia, Edited by Vedran Kordic, IN-TECH, 2009

Zhongfei (Mark) Zhang, Zhen Guo, Christos Faloutsos, Eric P. Xing, and Jia-Yu Pan, On the scalability and adaptability for multimodal image retrieval and image annotation, in Machine Learning Techniques for Adaptive Multimedia Retrieval: Technologies Applications and Perspectives, Edited by Roger Wei, Idea Group Inc., 2010

Bo Long, Zhongfei (Mark) Zhang, and Philip S. Yu, A General Framework for Relation Graph Clustering, Knowledge and Information Systems Journal, Elsevier Science Press, Accepted, 2009

Zhen Guo, Zhongfei (Mark) Zhang, Shenghuo Zhu, Yun Chi, and Yihong Gong, Knowledge Discovery from Citation Networks, Proc. IEEE International Conference on Data Mining, Miami, FL, USA, December, 2009

Zhen Guo, Shenghuo Zhu, Yun Chi, Zhongfei (Mark) Zhang, and Yihong Gong, A latent topic model for linked documents, Proc. ACM International Conference SIGIR, Boston, MA, USA, July, 2009

Xi Li, Weiming Hu, Zhongfei (Mark) Zhang, and Yang Liu, Spectral Graph Partitioning Based on A Random Walk Diffusion Similarity Measure, Proc. Asian Conference on Computer Vision, XiAn, China, September, 2009

Code Release :

Yi Xu, Zhongfei (Mark) Zhang, Bo Long, and Philip S. Yu, Pattern Change Discovery between High Dimensional Datasets, Proc. the 20th ACM Conference on Information and Knowledge Management, Glasgow, UK, October, 2011, (15% acceptance rate) [code]
Zhen Guo, Zhongfei (Mark) Zhang, Shenghuo Zhu, Yun Chi, and Yihong Gong, Knowledge Discovery from Citation Networks, Proc. IEEE International Conference on Data Mining, Miami, FL, USA, December, 2009 [code]

Data Release :

Yi Xu, Zhongfei (Mark) Zhang, Bo Long, and Philip S. Yu, Pattern Change Discovery between High Dimensional Datasets, Proc. the 20th ACM Conference on Information and Knowledge Management, Glasgow, UK, October, 2011, (15% acceptance rate) [data]
Zhen Guo, Zhongfei (Mark) Zhang, Shenghuo Zhu, Yun Chi, and Yihong Gong, Knowledge Discovery from Citation Networks, Proc. IEEE International Conference on Data Mining, Miami, FL, USA, December, 2009 [data]

This material is based upon the work supported by the National Science Foundation under Award No. 0812114.

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.

Go back to the Multimedia Computing Research Lab homepage