Optimized group formation for solving collaborative tasks

Many popular applications, such as collaborative document editing, sentence translation, or citizen science, resort to collaborative crowdsourcing, a special form of human-based computing, where, crowd workers with appropriate skills and expertise are required to form groups to solve complex tasks. While there has been extensive research on workers’ task assignment for traditional microtask-based crowdsourcing, they often ignore the critical aspect of collaboration. Central to any collaborative crowdsourcing process is the aspect of solving collaborative tasks that requires successful collaboration among the workers. Our formalism considers two main collaboration-related factors—affinity and upper critical mass—appropriately adapted from organizational science and social theories. Our contributions are threefold. First, we formalize the notion of collaboration among crowd workers and propose a comprehensive optimization model for task assignment in a collaborative crowdsourcing environment. Next, we study the hardness of the task assignment optimization problem and propose a series of efficient exact and approximation algorithms with provable theoretical guarantees. Finally, we present a detailed set of experimental results stemming from two real-world collaborative crowdsourcing application using Amazon Mechanical Turk.