Authoring Worked Examples for Java Programming with Human-AI Collaboration
Project Overview
The document explores the integration of generative AI in education through a human-AI collaborative approach, specifically focusing on the Worked Example Authoring Tool (WEAT) designed for Java programming. WEAT leverages large language models (LLMs) to generate detailed explanations for lines of code, addressing the challenges instructors face in creating high-quality educational content. A study conducted to assess the quality of AI-generated explanations against those crafted by experts reveals promising outcomes, indicating that generative AI can efficiently produce educational materials while potentially enhancing the learning experience for students. This approach not only streamlines the authoring process but also suggests that AI can serve as an effective tool for educators, fostering a more responsive and adaptable educational environment. Overall, the findings underscore the transformative potential of generative AI in supporting teaching and learning, particularly in technical subjects like programming.
Key Applications
Worked Example Authoring Tool (WEAT)
Context: Programming education for instructors and Teaching Assistants (TAs) in Java programming courses.
Implementation: Instructors provide code and problem statements; the AI generates initial explanations for review and editing.
Outcomes: Reduced authoring time for code explanations; improved quality of explanations as rated by users.
Challenges: Dependence on the quality of AI-generated explanations; potential need for instructor edits to ensure clarity and relevance.
Implementation Barriers
Technical
Quality of AI-generated explanations can vary, leading to the need for instructor intervention.
Proposed Solutions: Instructors can review and edit generated explanations to enhance quality and relevance.
User Experience
Instructors may struggle with prompt crafting for the AI to produce effective outputs.
Proposed Solutions: Providing clear guidelines and examples for prompt tuning to help users generate better responses from the AI.
Project Team
Mohammad Hassany
Researcher
Peter Brusilovsky
Researcher
Jiaze Ke
Researcher
Kamil Akhuseyinoglu
Researcher
Arun Balajiee Lekshmi Narayanan
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Mohammad Hassany, Peter Brusilovsky, Jiaze Ke, Kamil Akhuseyinoglu, Arun Balajiee Lekshmi Narayanan
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