A Game-Theoretic Approach to Quality Improvement in Crowdsourcing Tasks

Abstract

Together with the rise of social media and mobile computing, crowdsourcing increasingly is being relied on as a popular source of information. Crowdsourcing techniques can be employed to solve a wide range of problems, mainly Human Intelligence Tasks (HITs) which are easy to do for human, but difficult or even impossible for computers. However, the quality of crowdsourced information always has been an issue. Several methods have been proposed in order to increase the chance of receiving high quality contributions from the crowd. In this paper, we propose a novel approach to improve the quality of contributions in crowdsourcing tasks. We employ the game theory to motivate people towards providing information of higher quality levels. We also take into account players’ quality factors such as reputation score, expertise and the level of agreement between players of the game to ensure that the problem owner receives an outcome of an accepted quality level. Simulation results demonstrate the efficacy of our proposed approach in terms of improving quality of the contributions as well as the chance of successful completion of the games, in comparison with state-of-the-art similar methods.