Enhanced Question-Answering for Skill-based learning using Knowledge-based AI and Generative AI
Project Overview
The document explores the integration of generative AI with knowledge-based AI to improve skill-based learning in online education, highlighting the development of Ivy, an intelligent agent designed to enhance the learning experience. Utilizing a Task-Method-Knowledge (TMK) model in conjunction with large language models (LLMs), Ivy provides personalized and relevant explanations for skill-related queries, thereby fostering active engagement among learners. This innovative approach aims to bridge the gaps in traditional online education, particularly in the area of procedural knowledge, which often remains under-addressed. The findings indicate that by shifting the focus from passive learning to interactive and personalized experiences, generative AI can significantly enhance educational outcomes, empowering students to develop practical skills effectively. Overall, the document underscores the transformative potential of generative AI in creating a more engaging and effective online learning environment.
Key Applications
Ivy, an intelligent agent for skill-based learning
Context: Graduate-level online AI course at Georgia Institute of Technology
Implementation: Ivy uses a TMK model to provide detailed explanations to skill-based learning questions, utilizing generative AI for response generation.
Outcomes: Improved depth and relevance of feedback, fostering comprehensive understanding of skills crucial for problem-solving.
Challenges: Initial implementation as a standalone question-answering system; limitations in handling episodic queries.
Implementation Barriers
Technical barrier
LLM-based agents often provide general or shallow responses, struggling with deep procedural understanding.
Proposed Solutions: Integrating structured knowledge frameworks like TMK to enhance the explanatory capabilities of generative AI.
Developmental barrier
Manual creation of TMK models is time-consuming, requiring significant effort to ensure accuracy.
Proposed Solutions: Automate TMK model creation to reduce development time and improve scalability.
Project Team
Rahul K. Dass
Researcher
Rochan H. Madhusudhana
Researcher
Erin C. Deye
Researcher
Shashank Verma
Researcher
Timothy A. Bydlon
Researcher
Grace Brazil
Researcher
Ashok K. Goel
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Rahul K. Dass, Rochan H. Madhusudhana, Erin C. Deye, Shashank Verma, Timothy A. Bydlon, Grace Brazil, Ashok K. Goel
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