Personalized Cognitive Tutoring using Davinci-003 API for Adaptive Question Generation and Assessment
Project Overview
The document explores the innovative use of generative AI in education through a cognitive tutoring system that employs the Davinci-003 API to create personalized and adaptive questions for students, enhancing their learning experience and promoting knowledge transfer across diverse subject areas. This AI-driven tutor adjusts to individual student comprehension levels, offering tailored challenges designed to improve overall learning outcomes. A functional prototype was developed using Microsoft PowerApps, showcasing the practicality of integrating such technology in educational settings. Furthermore, the initiative emphasizes a commitment to educational equity, ensuring that all students have access to customized learning experiences that cater to their unique needs. Through these advancements, the document highlights the potential of generative AI to transform educational practices and foster more effective learning environments.
Key Applications
Cognitive tutor powered by Davinci-003 API
Context: Used in educational settings for personalized learning; target audience includes students of various backgrounds and learning needs.
Implementation: Developed a working prototype using Microsoft PowerApps that allows students to interact and receive personalized questions.
Outcomes: Improves student learning outcomes by providing adaptive questions that challenge students at their optimal difficulty level.
Challenges: Current limitations include lack of emotional analysis and inability to adjust question types or quantities based on individual preferences.
Implementation Barriers
Technical Challenge
The system cannot analyze emotional valence or tailor responses based on emotional states.
Proposed Solutions: Future research could integrate affective computing principles to adapt responses based on user emotional states.
Customization Limitations
The system lacks the ability to adjust the type or quantity of questions based on students' preferences or learning styles.
Proposed Solutions: Further development could focus on incorporating features that allow for more customization in question generation.
Project Team
Devan Walton
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Devan Walton
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai