Computer Science Department, PhD Dissertation Defense Sami Baral "Towards Effective Automated Feedback in Math Education"
9:00 a.m. to 10:00 a.m.
Sami Baral
PhD Candidate
WPI – Computer Science Department
Monday, April 28, 2025
Time: 9:00 a.m. – 10:00 a.m.
Location: Unity Hall 320
https://wpi.zoom.us/j/5088315569
Committee Members:
Advisor: Dr. Neil Heffernan, WPI - Computer Science
Dr. Lane Harrison, WPI - Computer Science
Dr. Xiaozhong Liu, WPI - Computer Science
Dr. Anthony F. Botelho, University of Florida - Educational Technology
Abstract:
Open-ended questions play an important role in mathematics education by encouraging critical thinking and fostering mathematical communication among students. However, due to the diverse nature of student responses, online learning platforms offer limited automated support for such questions. Teachers often have to manually evaluate open-ended responses, a process that is both time-consuming and delays feedback for students.
This dissertation addresses the critical need for timely feedback on open-ended responses by developing effective automated feedback systems. It explores advancements in automated scoring and feedback methods, leveraging traditional machine learning, natural language processing (NLP), and transformative large language models (LLMs) to enhance feedback strategies on online learning platforms. While LLMs show significant potential for automating feedback, there is limited evidence of their effectiveness in real classroom settings, leaving questions about their impact on learning outcomes unanswered.
To bridge these gaps, this dissertation presents: (a) an evaluation process that involves input from educators to assess the quality of LLM-generated feedback, and (b) a study to measure the effects of immediate automated feedback on student learning outcomes. By incorporating input from educators and conducting empirical studies with students, this work aims to improve feedback methodologies in online learning platforms