Large-Scale Online Collaboration in Hierarchical Networks
Abstract
Statement of Work:Prof. Christoph Riedl will be directing the project at Northeastern University and will contribute to the work on the (1) development of formal mathematical and agent-based models to identify and locate inefficiencies; (2) development and implementation of an experimental framework; (3) design and execution of human subject experiments; (4) the data analysis activities from experiments; (5) development and evaluation of (social) interventions. He will supervise the work and training of the post-doctoral fellow and graduate student at Northeastern. He will participate in the various aims of the project concerning the development and implementation of the experimentation platform. He will contribute to the developing of experimentation and measurement framework and analytical and computational techniques for data analysis. Prof. Riedl will supervise the work concerning the development of large-scale experimentation approaches and their integration into a common platform. He will supervise the ongoing work regarding the development of formal and agent-based models, their simulation, analysis and interpretation.Objective:The goal of this research is to (1) advance our understanding of inefficiencies in decision-making in large organizations and (2) propose interventions on how to mitigate their negative impact. In order to achieve this goal, the project employs an integrated approach of modeling and experimentation that blends theories and methods from social science with computational methods from computer science and complex systems. The research approach comprises four objectives:1D. evelop formal mathematical and agent-based models to identify and locate inefficiencies.2D. evelop and implement an experimental framework that can be used to test and calibrate the models.3C. onduct human subject experiments using the experimental framework developed in (2) in an attempt to empirically demonstrate and validate the inefficiencies identified by the model and to calibrate parameters (iterate with modeling). 4D. evelop and evaluate (social) interventions that could be used to alleviate (some) inefficiencies.Approach:This projects applies (and seeks to further develop) a new paradigm of social science research that focuses onidentification of causal mechanisms that allow the design of massively scalable interventions. The processincorporates feedback from iteration of modeling, experimentation, and testing of implemented control solutions: (1) predictive modeling to identify actionable bottlenecks (e.g., using agent-based modeling); (2) human-subject experimentation to suggest scalable interventions; and (3) design, implement, and evaluate scalable interventions to demonstrate replication and expand interpretation. This integrated and iterative research processes allows that results from each step critically inform each other. An important innovation and key strength of this research lies in the combination of empirical experimentation and formal modeling. Agent-based modeling will be performed by building ona existing framework developed by the PI and his team using scalable Matlab code and the high-performancecomputing infrastructure available at Northeastern University hosted at the Massachusetts Green High Performance Computing Center. Human Subject Experimentation will leverage open source materials and an existing online experimental platform developed and hosted at Northeastern University called Volunteer Science. Volunteer Science is a collaboration among scientists from leading research universities to expand the tools available for social and behavioral research.ONR Mission/Relevance:Today, nations face an uncertain and complex security landscape. Precise decisions require a high amount ofinformation processing and the evaluation of potential solutions. Complex knowledge needs to be integrated and combined including diplomatic, economic, intelligence, and military information. On a general level, collaborati
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Sep 23, 2016
- Source ID
- N000141613005
Entities
People
- Christopher Riedl
Organizations
- Northeastern University
- Office of Naval Research
- United States Navy