Advancing Human-based Modeling with Large-Scale High Performance Computing
Abstract
Social computing has become one of the great frontiers of 21st Century science. New tools, theories and techniquesare growing rapid""ly to collect, quantify, model and predict massive data sets containing diverse and detailedinformation about human activities. Our" team has become leaders in developing innovative ways to analyze large scalehuman systems ~ currently we are PIs on 12 funded DoD" grants which use social computing methods. Collectively,our research has contributed widely to our understanding of social process""es, identifying mechanisms and fundamental principles underlying team performance, collective decision-making, group and cultural co""nsensus, network resilience, communication and mobility patterns, disease and opinion dynamics, and community activism. Despite achi""evements in this field, we do not yet have a centralized computationalpower to facilitate major discoveries. Over the last five yea""rs, Northeastern University has built an infrastructure upon which we are now able to leverage and expand enabling us to purchase ou""r own nodes adding to universitywide high performance computing cluster. Ourproposal includes a centralized, state-of-the-art, secu"re computing cluster dedicated for research and education of the Network Science Institute faculty and students. We propose to build" four dedicated computer sub-clusters: two CPUbased computing queues (standard and large memory), one Hadoop queue, and one Nvidia T""esla-based GPU computing queue. As the core faculty of the Network Science Institute, we represent diverse expertise, with equally d""iverse projects, and supporting the only PhD program in Network Science in the US. A state-of-the-art computing environment is criti"cal for our research and to train the next generation of thought leaders to understand how to leverage massive data sets of human ac"tivity to better predict, influence, and design social infrastructures. Thus, the instrumentation granted by the DURIP program would"" contribute to the development of a growing community of big data, network science scholars.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- May 05, 2017
- Source ID
- N000141712542
Entities
People
- Kathryn Coronges
Organizations
- Northeastern University
- Office of Naval Research
- United States Navy