MRI/Acq.: FUTURO: A Data Intensive and High Performance Computing Cluster for Integrated R & E
This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).
Project Proposed: This project, acquiring Futuro, a computer cluster for interdisciplinary research projects and Computer Science education programs, enables research activities in data mining, pattern discovery, genetic data analysis, experimental astronomical physical, collaborative filtering, theory of computation, high dimensional visualization, and other computational areas. Seeking to lay the foundation for a strong research and education integration centered on CS in two minority-serving universities, UT-Brownsville and UT-Pan American, it enables the following potential transformative goals: - Terabyte scale data mining and pattern discovery in time series datasets obtained from heterogeneous sensor networks (addresses data analysis problems in Laser Interferometer Gravitational Wave Observatory (LIGO)) - Genetic data analysis in complex human diseases to identify susceptibility factors enabling understanding of genetic causes of complex diseases such as schizophrenia (potential to lead to new therapeutic strategies) - Studying the dynamical systems and Stellar populations to model the behavior of black hole binaries in globular cluster and galactic nuclei (creates models of formation of stellar systems via intensive computation that can provide information for interpretation of results from operating gravitational wave detectors) - Exploring and creating computing-effective, scalable, robust and intelligent learning algorithms for large recommender systems based on collaborative filtering by incorporating multispectral information (may lead to next generation recommender systems) - Visualizing high dimensional streaming data from heterogeneous sensors (potential to contribute in developing new data reduction methodologies that incorporate intelligent computation such as data mining and thus more advanced visualization systems with cross-disciplinary utility) - Benchmarking and developing algorithms for approximating NP-hard subgraph isomorphism problem with best possible practical performance (benefits applications such as image recognition and bioinformatics) Futuro will serve as laboratory in which core research can be conducted in a collaborative fashion at a high level providing real-world test applications while training students.
Broader Impacts: This project benefits many users from physics, bioinformatics, computational engineering, and environmental engineering in two minority-serving universities. Futuro forms the nucleus for collaboration between computer scientists and researchers from other departments. The project will train students at the Rio Grande Valley, a historically underrepresented region with more than 90% Hispanics, in areas that are expected to have great national impact. The work provides experience in parallel programming and scientific computing, the CS curriculum will be enriched by the lab modules enabled by the cluster facility.
