The Information Management and System Group is currently conducting research on next-generation multimedia management technologies, with the mission to realize the Microsoft .NET vision for pervasive media management on the Internet. Our goal is to bring the Internet and the user's experience with multimedia to the next level by developing intelligent media analysis algorithms and system and network technologies to make multimedia management part of the Internet infrastructure services easily accessible to anyone, anywhere, anytime, and through any type of device.

Smart content technology and adaptive content delivery 
In the PC+ era, new computing devices with diverse capabilities are making a population boom. Document-related applications are facing a big challenge C the problem of multiple form factors. In this project, we devote ourselves to (a) developing of content representation which is scalable and adaptable, (b) developing of content analysis techniques to structure content for the new representation, and (c) developing of corresponding rendering algorithms and innovative user interfaces to maximize the information throughput of the content on any devices.


Web Search: 
The research of Web search becomes more and more important with the rapidly explosion of the entire Web information. Our goal is try to help MSN as the best search engines in the World. In this project, we will put our efforts on several aspects (not restricted to followings): (1) developing of large scale Web search platform and evaluation platform, (2) Enhanced page parser (3) utilizing text mining techniques to enhance Web categorization and clustering (4) Connecting search with business by paid search (5) explore vertical search, including newsgroup search, community search, help& support search, news search, etc. (6) Improve search results by relevance measurement (7) personalized search results by analyzing users' search behaviors


Text Mining and Knowledge Management  
With the explosive growth of textual information in the World Wide Web and enterprises, we are currently drowning in "data oceans" and facing serious "data overload". Current information seeking tools, mainly based on traditional IR technologies, are insufficient to truly meet users' information needs. We foresee that the biggest challenge in the next several decades is how to effectively and efficiently dig out a machine-understandable information and knowledge layer from unstructured or semi-structured text data. Therefore, the main goal of this project is to discover and organize information and knowledge hidden in texts with various types of formats, thus substantially improving information acquisition, sharing and searching.

Text mining and knowledge management is a very broad space, and it requires many techniques from basic research areas including information extraction, information retrieval, machine learning, data mining and natural language processing. Especially, we are exploiting to develop effective and scalable mining approaches in the following topics:
1. Extracting knowledge from enterprises' troubleshooting database to improving the productivity of technical support
2. Mining newsgroup data to facilitate advanced search and management
3. Mining deep websites for information integration and deep web search
4. Community mining
5. Linguistic analysis of web documents

P2P & Distribured system:
One of the most exciting research opportunities in the system research is the self-organized P2P systems. Work in this space combines good principles of distributed system research as well as cues from other disciplines such statistics, economics and sociologies.

We are conducting basic infrastructure research on high-performance, robust P2P distributed hash table (XRing), in-system self-organizing monitoring service (SOMO), and fundamental primitives such as highly-available distributed mutual exclusion protocol (the Sigma protocol). These basic research works will have profound impact on many important applications that we are developing in parallel, including a self-administrated, self-tuned highly available storage system (RepStore), a wide-area P2P resource pool (ImagineONE.net lab), wide-area application-level multicasting. Our end vision is a self-organizing and self-evolving next-generation distributed operating system.

MiXP: A personalized and intelligent media search service:
In order to help end-users effectively and efficiently manage their personal media files, we are developing MiXP, which is an intelligent web service that are able to automatically collect and build personalized semantic indices of media files on behalf of end-users. MiXP provides end-users a single, unified control point and easy access and management of their personal media files from any of their devices. MiXP also learns form the users usage patterns and interactions to refine the indices and to model the users intentions and preferences so as to provide higher quality services. With the users permission, MiXP may serve as the users delegate to interact with other Microsoft .NET based Web applications to provide personalized services.

MediaLand: A universal platform for multimedia database:
The goal of MediaLand is to develop the database platform for managing multimedia data and their entire lifecycle (including data modeling, storing, indexing, querying and searching) by leveraging existing data management techniques. We are developing a uniform language (and GUI) to express users query requirements. It has an intelligent mediator which refines user queries and gets final query execution plan by choosing best approach to search the most appropriate data sources. We are also performing research on comprehensive query and search techniques to support hybrid information requirements for complex queries.

• Patrick Baudisch, Xing Xie, Chong Wang, Wei-Ying Ma, Collapse-to-Zoom: Viewing Web Pages on Small Screen Devices by Interactively Removing Irrelevant Content,17th Annual ACM Symposium on User Interface Software and Technology (UIST 2004), TechNote, Sante Fe, NM, Oct. 2004.
• Xin Zheng, Deng Cai, Xiaofei He, Wei-Ying Ma and Xueyin Lin, Locality Preserving Clustering for Image Database ,12th ACM International Conference on Multimedia, New York City, USA, Oct. 2004.
• Deng Cai, Xiaofei He, Zhiwei Li, Wei-Ying Ma and Ji-Rong Wen, Hierarchical Clustering of WWW Image Search Results Using Visual, Textual and Link Analysis ,12th ACM International Conference on Multimedia, New York City, USA, Oct. 2004 .
• Xiaofei He, Wei-Ying Ma, Hong-Jiang Zhang, Learning an Image Manifold for Retrieval,12th ACM International Conference on Multimedia, New York City, USA, Oct. 2004.
• 13. Xin-Jing Wang, Wei-Ying Ma, Gui-Rong Xue, and Xing Li, Multi-Model Similarity Propagation and its Application for Web Image Retrieval,12th ACM International Conference on Multimedia, New York City, USA, Oct. 2004.
• Jun Yan, Benyu Zhang, Shuicheng Yan, Zheng Chen, Weiguo Fan, Wensi Xi, Qiang Yang, Wei-Ying Ma, and Qiansheng Cheng IMMC: Incremental Maximum Margin Criterion,10th ACM SIGKDD international conference on Knowledge discovery and data mining, Seattle, USA, Aug. 2004.
• Jiying Wang, Ji-Rong Wen, Fred Lochovsky and Wei-Ying Ma, Instance-based Schema Matching for Web Databases by Domain-specific Query Probing, The 30th International Conference on Very Large Data Bases (VLDB 2004), Toronto, Ontario, Canada, August 2004.
• Dou Shen, Zheng Chen, Hua-Jun Zeng, Benyu Zhang, Qiang Yang, Wei-Ying Ma, Yuchang Lu, Web-page Classification through Summarization, The 27th Annual International ACM SIGIR Conference (SIGIR'2004), July 2004.
• Hua-Jun Zeng, Qi-Cai He, Zheng Chen, Wei-Ying Ma. Learning To Cluster Search Results, The 27th Annual International ACM SIGIR Conference (SIGIR'2004), July 2004.
• Ji-Rong Wen, Ni Lao and Wei-Ying Ma, Probabilistic Model for Contextual Retrieval,The 27th Annual International ACM SIGIR Conference (SIGIR 2004), July 2004 .
• Deng Cai, Shipeng Yu, Ji-Rong Wen and Wei-Ying Ma, Block-based Web Search, The 27th Annual International ACM SIGIR Conference (SIGIR 2004), July 2004 .
• Deng Cai, Xiaofei He, Ji-Rong Wen and Wei-Ying Ma, Block-Level Link Analysis,The 27th Annual International ACM SIGIR Conference (SIGIR 2004), July 2004 .
• Xiaofei He, Deng Cai, Haifeng Liu and Wei-Ying Ma. Locality Preserving Indexing for Document Representation,The 27th Annual International ACM SIGIR Conference (SIGIR'2004), July 2004.
• Deng Cai, Xiaofei He, Wei-Ying Ma, Ji-Rong Wen and Hong-Jiang Zhang, Organizing WWW Images Based on the Analysis of Page Layout and Web Link Structure,2004 IEEE International Conference on Multimedia and Expo., Taipei, Jun. 2004.
• Xing Xie, Wei-Ying Ma, Hong-Jiang Zhang, Maximizing Information Throughput for Multimedia Browsing on Small Displays,2004 IEEE International Conference on Multimedia and Expo., Taipei, Jun. 2004.
• Yusuo Hu, Xing Xie, Zonghai Chen, Wei-Ying Ma, Attention Model Based Progressive Image Transmission,2004 IEEE International Conference on Multimedia and Expo.Taipei,Jun. 2004
• Xin-Jing Wang, Wei-Ying Ma, and Xing Li, Data-Driven Approach for Bridging the Cognitive Gap in Image Retrieval, 2004 IEEE International Conference on Multimedia and Expo., Taipei, Jun. 2004.
• Ying Liu, Xiaofang Zhou, Wei-Ying Ma, Extracting Texture Features from Arbitrary-shaped Regions for Image Retrieval,2004 IEEE International Conference on Multimedia and Expo., Taipei, Jun. 2004.
• Ruihua Song, Haifeng Liu, Ji-Rong Wen and Wei-Ying Ma, Learning Block Importance Models for Web Pages,The Thirteenth World Wide Web conference (WWW 2004), 203-211, New York, May, 2004.
• Wensi Xi, Benyu Zhang, Yizhou Lu, Zheng Chen, Shuicheng Yan, Huajun Zeng, Wei-Ying Ma, and Edward A. Fox. Link Fusion: A Unified Link Analysis Framework for Multi-Type Interrelated Data Objects ,The Thirteenth World Wide Web conference (WWW 2004), 203-211, New York, May, 2004.