Preliminary Algorithmic Foundations for Ranking Quizzes and Students from Student-sourced Quizzes

点击翻译按钮获取中文摘要

The project addresses a new set of problems that will increasinglyarise with a gradual shift to massively open online courses andlarge classes where new assessments to test active learning of theconcepts are costly to design. The context of the project is aprototype cloud-based quiz management system that leads the onlinelearners through scaffolded activities, such as creating,answering, and improving of student-sourced quiz items. The systemthen automatically evaluates the competence of learners and thequality of questions from outcomes of these activities. Thisproject addresses the key algorithmic issue in such systems ofefficiently determining the quality of the students and thesubmitted items. Preliminary ideas for new algorithms will bedeveloped and deployed in the test system to understand theiradvantages and shortcomings. The development and testing ofalgorithms will provide opportunities for training two graduatestudents. This will also lead to a refinement of the eventualalgorithmic design of a system for scalable online assessments withsignificant societal impacts.In particular, the project develops new algorithmic frameworks forproblems in the joint ranking of students and student-sourced quizquestions based on feedback in the form of peer performance andreview. A key contribution is the formulation of these problems interms of abstract problems amenable to both combinatorial andcontinuous optimization, which in turn allows a completelyunrelated subfield of people (those of optimization) to work onthese problems in the future. At the same time, the project willundertake the groundwork of developing the first such algorithmsthat solve these problems with current approaches.