Abstract: Detecting, analyzing and reasoning anomalies is important for many real-life application domains such as computer networking, fraud analysis, and software security. The main challenges include the overwhelming number of low-risk events and their multifaceted relationships, the diversity of anomalies by various data and anomaly types, and the difficulty to incorporate domain knowledge in the anomaly analysis process. In this talk, we introduce a suite of novel concepts called correlation graphs (CG). CG achieves computational scalability, domain generality, and user interactivity through synthesizing heterogeneous types of objects, their anomalies, and multifaceted relationships in a single graph. We propose relevant anomaly detection algorithms, elaborate visualization designs, and interaction models to allow users to fully unleash the visual analytics power over the correlation graph. Case studies in real-world applications are presented that demonstrate the effectiveness of CG in the data-driven detection, visualization, and reasoning process of anomalies.

 

Bio: Lei Shi is a professor in the School of Computer Science and Engineering, Beihang University. Before that, he was a professor in SKLCS, Institute of Software, Chinese Academy of Science, and a research manager at IBM Research - China. He holds B.S. (2003), M.S. (2006) and Ph.D. (2008) degrees from Department of Computer Science and Technology, Tsinghua University. His current research interests are Visual Analytics and Data Mining, with more than 80 papers published in top-tier venues, such as IEEE TVCG, TKDE, TC, VIS, ICDE, Infocom, ACM Sigcomm and CSCW. He is the recipient of IBM Research Division Award on "Visual Analytics" and the IEEE VAST Challenge Award twice in 2010 and 2012. He has organized several workshops on combining visual analytics and data mining, in ICDM, CIKM, etc, and served on the program committees of many related conferences such as VIS, KDD, IJCAI, AAAI, Graph Drawing, etc. He is an IEEE senior member.