Big data technology is widely adopted across many disciplines. Big data is complex, volatile, lack of correlation, and value scarce by nature, which makes it difficult to form standardized and systematic technological solutions, to address the diversified requirements for life cycle management of big data in different application domain. In order to build sustainable big data application systems, encourage its rapid development and delivery of expected values with minimum efforts, we development innovative engineering technology and integrated platform for big data applications. Major challenges to be addressed include: data life cycle management covering: data collection, storage, computation, analysis, visualization, as well as the software systems engineering life cycle.
Event data management focuses on the management, mining and analysis of massive amount of event data. Major data types include temporal data, graph data processing algorithms, system design and applications. Key technologies include: event data storage, feature identification, feature indexing, efficient searching, etc. We have published 80 + papers in this area at top level conferences and journals, e.g., SIGMOD, VLDB, ICDE, IEEE TSC, WWW, Computers in Industry, DMKD, JIIS, SoSyM、Information Processing Letters, referenced for 1000+ times from Google Scholar.
Media Data search and analysis focuses on the area of multimedia information retrieval and management, in particular, visual object classification, automatic semantic annotation, content-based multimedia indexing, social multimedia retrieval, mining and recommendation. The media group has published more than 50 research papers in international conferences and journals (CVPR, ICCV, SIGIR, ICML, AAAI, IEEE TKDE, ACM TIST, CVIU, MTAP) and applied for 8 Patent Rights in China.
Machine learning is a well-recognized area, which is to design a computational process of acquiring new knowledge or skills, and optimizing system performance by getting inspirations from human behaviour. The web data mining group’s research interests include machine learning and data mining algorithms and techniques of unstructured web data, social networks, graph data, stream data. The group has published over 100 papers on relevant conferences and journals, including TKDE、PVLDB、KDD、SIGIR、IJCAI、AAAI、ACM Multimedia, KDD, IJCAI, WSDM, CIKM, ADKDD, etc.
Today’s industrial processes are tightly coupled with information technology, the product data accumulated daily has far beyond the processing ability of conventional data processing approaches. Industrial Big Data group aims to provide key enabling technology for enterprises to build new business services using accumulated data, in particular, the data storage, process, mining and generation of added value to make the transformation of traditional industrial organizations into advanced high-end manufacturing and service enterprises. In due process, the data objects are mostly sensor data produced by construction machinery. Key techniques for big data management and mining are adopted to gain patterns, experiences, insights and value from data.
The recent rapid development of healthcare information technology leads to the accumulation of hospital operational data and patient clinical records. By using knowledge management and big data analytics techniques, we aims to identify general patterns and gain insights from the available data resources, so that value-added services can be provided to patients health management, medical practitioners’ clinical decision making, and healthcare administrator’s clinical quality measurements.