CAREER: Enabling expert crowdsourcing via coordination, targeted contribution and education

Stanford University, Stanford CA

Investigators

Abstract

Crowdsourcing systems typically draw on the work of non-experts, as expert crowds are difficult to gather and coordinate. To extend the reach of crowdsourcing, this project investigates interactive systems, platforms and computational techniques to integrate experts as core participants of crowdsourcing systems. The project will develop and evaluate three related systems that improve the coordination, participation and education of expert crowds. First, the project will develop modular, composable workflows to guide paid expert crowds to accomplish complex tasks such as design and engineering. A second system seeks to attract a new form of expert participation by reaching out to experts in their spare moments. A final system will enable crowd workers to learn new skills by leveraging existing tasks as work-study opportunities. This research will produce three concrete types of results: 1) techniques and patterns for guiding expert crowds and their contributions, 2) scientific results and evaluations that depict the strengths and weaknesses of expert crowds, and 3) open, public platforms and systems to recruit, guide, and train members of expert crowds. Crowd work has the potential to employ millions of full-time workers and grant them the flexibility to guide their own careers. The proposed work advances a vision of how experts can engage in the future crowd-work economy, contribute to projects while improving their own skills, and be supported by practical knowledge and new sociotechnical systems. The project also includes a plan to use expert work-study techniques to provide realistic training in Stanford?s HCI curricula and develop a peer mentoring system to scale advising and informal learning.