Core Concepts
Designing a personalized tutoring system with student modeling involves diagnostic components and LLM prompt-based tutoring to enhance learning outcomes.
Abstract
The content discusses the challenges and design considerations for creating a personalized tutoring system using Large Language Models (LLMs) in conversation-based education. It outlines the key components of student assessment, adaptive exercise selection, and prompt design for personalized tutoring. The results of the experimental analysis on adaptive exercise selection and learning gain are also presented.
-
Introduction
- Educators' interest in leveraging Large Language Models (LLMs) for personalized tutoring systems.
- Challenges in accurately assessing students and incorporating assessments into teaching.
-
Design of Personalized Tutoring System
- Student Assessment: Cognitive state, affective state, learning style.
- LLM Prompt-Based Personalized Tutoring: Adaptive exercise selection, prompt design.
-
Results
- Adaptive Exercise Selection: Average correctness ratio of exercises presented during tutoring sessions.
- Learning Gain: Calculation of learning gains based on pre-test and post-test proficiency levels.
-
Experimental Results
- Detailed specifications on adaptive exercise selection and learning gain calculations.
-
Conclusion
- Identification of areas for improvement in connecting student assessments to effective tutoring strategies and enhancing user engagement.
Stats
"The dataset was utilized to train the IRT model and establish the item parameters required for calculating the student’s proficiency."
"The average Area Under the Receiving Operator Curve (AUC) was 0.65."
"For all 20 participants, the average learning gain was calculated as −0.0753, 0.0159, and −0.0102 in Pronouns, Punctuation, and Transitions."
Quotes
"It’s interesting to note that based on your pre-test results, I see that your knowledge about punctuation is actually stronger than you think! That’s a great start."
"Encouraging the student to go beyond the right answer and explain the reasoning behind their choices in their own words proved to be an effective strategy."