While long-lived communities have been one of the key organizing principles of Web-based systems, there is widespread evidence of highly-dynamic, ad-hoc crowd formation in emerging real-time socio-computational systems. These crowds are dynamically formed and potentially short-lived, often with only implicit signals of their formation and evolution. The goal of this research project is to develop the framework, algorithms, and systems for lightweight crowd-oriented search and computation so that stakeholders can distill high-quality information from bursty social systems and actively engage with the crowds generating this information. First, the project provides the foundation for crowd-oriented search through new algorithmic advances for distributed crowd indexing and in an investigation of the design principles impacting crowd-oriented search. Next, the project develops self-tuning methods for assessing crowd quality, even with huge demands on efficiency and in the presence of limited evidence of crowd quality. Finally, the project explores methods for "closing the loop" in crowd-oriented search, so that crowds may become part of in situ human-computational systems.
The education and outreach efforts of the project are tightly linked to the research goals through leadership workshops, enhancements to the curricula, direct research training, and engagement with emergency response experts and major companies. Distilling high-quality information from bursty social systems and actively engaging with the crowds generating this information will result in improved real-time decision-making, impacting a wide range of stakeholders from areas such as epidemiology, law enforcement, government, finance, politics, among many others. Further information can be found on the project web page: http://faculty.cse.tamu.edu/caverlee/csc/.
