Tracking Human Process Using Crowd Collaboration to Enrich Data

Abstract

A rich source of data that has been largely ignored in crowdsourcing is the processes that humans use to accomplish a task. If we can capture this information, we could model it for automatic processing and use it to better understand the phenomena being modeled. Using crowd collaboration to trace workers' process produces a rich dataset that can be mined for new insights. We tested this approach on the task of sentence simplification and show the sources and types of additional information that we have obtained.