Stream Data Analytics

A major aspect of Big Data is velocity, namely data are being generated in fast speed from various sources such as network, sensor, etc. How to conduct analytics on these streaming data is a core challenge in Big Data. Research problems we are working on are : 1) stream data analytics as a service and its optimization; 2) real-time intrusion detection from streaming network data; 3) imbalanced stream data mining.

Big Data + High Performance Computing + Atmospheric Sciences

Besides theory and experimentation, computation has become the third pillar and data-driven science has become the fourth pillar of the scientific discovery process for many disciplines and critical to their research advances, such as bioinformatics, computational chemistry, computational physics and mechanical engineering. Collaborating with researchers in Applied Mathematics and Physics, we are conducting multidisciplinary research and education that require both high performance computing (HPC) and Big Data. The current scientific focuses are Atmospheric Science challenges including modeling and analysis of atmospheric radiation budget and evaluation of numerical global climate models (GCM). More at http://cybertraining.umbc.edu/.

Smart Manufacturing

Partnering with DOE Clean Energy Smart Manufacturing Innovation Institute (CESMII) and Smart Manufacturing Leadership Coalition (SMLC), we are exploring techniques to improve productivity, reduce cost and energy in manufacturing: 1) workflow as a service (WFaaS) to seamlessly integrate analytics pipelines in manufacturing and provision them as Cloud services; 2) sensor data analytics based anormaly detection and predictive maintenance in manufacturing; 3) service composition based agile manufacturing.