By April Groce
Faculty Mentor: Prashant Chandrasekar
Abstract
Project-based learning (PBL) provides students with a valuable, primarily self-guided learning experience. However, the instructor still plays a crucial role in guiding and assessing teams, which can be challenging in such an environment. Assessing PBL can be demanding due to the wide array of disciplines being addressed in the projects. Additionally, facilitators of PBL have limited time, so providing specialized, helpful feedback for each project can be difficult. A categorization system can be used for PBL activity flow as a method of identifying a current team’s progress towards their goal. My work proposes an artificial intelligence system that categorizes sentences from weekly reports in a PBL setting. To use this AI system for guidance of future action, it is important that the instructor understands why the system predicted the category that it did. My work also includes the use of Local Interpretable Model-Agnostic Explanations (LIME) as a method of explaining an AI model that is used to influence future action.
Leave a Reply