Supporting Complex Work in Crowdsourcing Platforms: A View from Service-Oriented Computing

Crowd sourcing is changing the way people work and solve problems from "in-house working" to "public out-sourcing". Most online crowd sourcing platforms perform two main functions: (i) allowing users to advertise their tasks and (ii) helping them find candidate workers. However, they do not support crowd sourcing of complex work consisting of interdependent tasks. Those tasks require not only having simple task/worker pairs, but also coordinating multiple workers together for completion of the crowd work. In this paper, we propose a conceptual framework to bring the coordination support into current online crowd sourcing platforms. In our framework, each crowd worker is modeled as a service that can be self-described, dynamically discovered and assembled into the complex crowd work, meanwhile, we define a workflow-based schema to structure and represent the complex crowd work. Then the crowd sourcing performance is initiated and managed by our coordination protocol.