Adaptation and incentive design in large scale task systems (doctoral consortium)

My thesis focuses on understanding and influencing the behavior and performance of users in large scale task systems. In task systems, multiple participants work independently to solve problems that are presented to them by a computer. From e-learning systems like Khan Academy and Coursera, to citizen science systems like Zooniverse and BugGuide, large scale task oriented systems are booming on the Internet and are growing in number and in size. These systems provide a gold mine of data, containing information about the ways human beings engage with tasks of various natures. In my work I tackle three challenges which are central to the design of successful task systems: (1) understanding and modeling users' behavior in large scale task systems (2) designing incentive mechanisms for influencing users' behavior and improving their performance (3) adapting interaction in task systems to the needs and strengths of individual users. My research activities combine computational models, algorithms and empirical methodologies to meet the challenges above. They are conducted in the context of two different types of large scale task domains consisting of e-learning systems and citizen science systems. I will evaluate my approach in real world environments. My preliminary results demonstrate my proposed approach by significantly outperforming the state-of-the-art methods in personalization of tasks to students in two separate e-learning datasets.