The members of the Informatics Group perform research and development in the broad area of information management and analysis. Research topics include the development of new algorithms in bioinformatics, data management for large-scale attributed graphs, reconfigurable hardware acceleration for data-intensive computing applications, novel approaches for analyzing large collections of text, analysis and understanding of dynamic networks, streaming algorithms, and cyber security. Many of our problems involve very large data sets, such as text collections of tens of millions of documents, graphs with billions of edges, or streaming cyber data at hardware line speeds. We typically employ a variety of technologies and tools, such as machine learning and classification algorithms found in packages like WEKA, hardware such as the Tilera multicore engine, streaming middleware such as IBM’s Infosphere Streams, and open-source tools such as Hadoop and SOLR.
The customers of the Informatics Group include scientists and analysts both at the Laboratory and in the US Government. Many of our members work with our Global Security Directorate, and we receive external funding from the DOE Office of Science, DARPA, and various other U.S. Government agencies.
Brian Van Essen: spatial accelerators for embedded systems and HPC, reconfigurable computing, and memory architectures for data-intensive computing
Peter Barnes: parallel discrete event simulation tools and applications, communication networks
David Buttler: database technology, web data access, web service selection, web document change detection
Brian Gunney: adaptive mesh refinement, HPC, numerical methods for partial differential equations, computational physics
Keita Iwabuchi: large-scale graph processing, out-of-core processing, HPC
Chandrika Kamath: data mining, machine learning, signal and image processing, HPC
Scott Kohn: cyber security, graph data management and analysis, HPC
Celeste Matarazzo: cyber security, distributed agents, distributed sensor analysis
Roger Pearce: distributed file systems and tools to profile the I/O performance of data intensive applications
Jae-Seung Yeom: data-dependent application behavior modeling, performance analysis, epidemic, viral evolution simulations
Andy Yoo: scalable large graph algorithms, large-scale data management, data mining and knowledge discovery, high performance parallel computing, parallel algorithms