NExT is an acronym for “NUS-Extreme-Tsinghua”

It is a center jointly setup between National University of Singapore (NUS) and Tsinghua University with the support of Media Development Authority (MDA) of Singapore. The term extreme search refers to the ability to search for live and dynamic data beyond what is indexed in the Web.

The center focuses on crawling and mining user-generated contents (UGC’s) in Singapore and Beijing in the following areas:

1. location-oriented information: includes shared photos and check-in venues;
2. topic-oriented information: covers forums, question-answering, tweets;
3. application-oriented information: covers mobile applications and associated information and discussions; and
4. structured information: includes factual, cultural and historical information.

These are publicly available data sources documenting the social interactions of people within a city. It will cover social activities of people and their collective preferences and interests. This gives rise to a graph of social pulses, and the research is focusing on transforming the unstructured, multisource and multimodal UGC contents into the graph depicting the social pulses of a city. The Center will work closely with other MDA funded centers to integrate their resources such as the sensor data streams (from CSI), language translation and resources (from CSIDM), as well as various MDA start-up companies and government agencies to aggregate related structured contents. The Center will offer the organized information base, as well as related information processing tools, available to other research entities, academic institutes and start-up companies to develop innovative applications to support the users.

Research Focuses

Many research efforts have been done on the analysis, indexing, fusion and retrieval of Web, deep Web and multimedia contents. This centre focuses instead on the novel and more challenging topic of handling live information sources. To make these live and dynamic data available for general access, we need to tackle the following research issues.

1. Develop distributed and database infrastructures to capture, analyze, and index the huge amount of continuous live data streams and their key analytics.
2. Bridge the semantic gaps of understanding the contents of non-textual data arising from video, audio and other type of sensors.
3. Build tools to analyze the contents of media sensors, the multilingual and often ungrammatical Blog and forum postings, and information generated from vast number of mobile devices.
4. Fuse these live information sources to extract the latest events and understand the current hot topics of discussion and their sentiment, as well as the relationships among users and between users and resources.
5. Build mobile platform to realize a conducive, informed environment to help users lead better life. Initial platforms are planned for campus and health care domain.