Speaker: Swaprava Nath, post-doctoral researcher at Indian Statistical Institute, New Delhi, India.

Abstract: Organically grown crowdsourcing networks, which includes production firms and social network based crowdsourcing applications, tend to have a hierarchical structure. Considering the entire crowdsourcing system as a consolidated organization, a primary goal of a designer is to maximize the net productive output of this hierarchy using reward sharing as an incentive tool. Every individual in a hierarchy has a limited amount of effort that they can split between production and communication. Productive effort yields an agent a direct payoff, while the communication effort of an agent improves the productivity of other agents in her subtree. To understand how the net output of the crowdsourcing network is influenced by these components, we develop a game theoretic model that helps explain how the individuals trade off these two components depending on their position in the hierarchy and their shares of reward. We provide a detailed analysis of the Nash equilibrium efforts and a design recipe of the reward sharing scheme that maximizes the net productive output. Our results show that even under strategic behavior of the agents, it is sometimes possible to achieve the optimal output and also provide bounds on the achievability when this is not the case.

More information on XRCE seminars.