Experimental Interface for Multimodal and Large Language Model Based Explanations of Educational Recommender Systems
Project Overview
The document explores the implementation of a web-based tool that leverages generative AI to improve educational recommender systems by offering multimodal explanations to enhance learner understanding and acceptance. This tool integrates various modalities, such as text, visual aids, and chatbot interactions, to provide comprehensive explanations of AI-based recommendations. Initial evaluations show promising results, indicating high levels of user acceptance and increased motivation among learners to engage with the recommended content. However, the findings also highlight the need for further research to overcome the limitations associated with large language models (LLMs) in educational settings. Overall, the use of generative AI in education demonstrates potential for enriching the learning experience, although challenges remain that require ongoing investigation to ensure effective implementation.
Key Applications
Web-based tool for multimodal explanations of educational recommendations
Context: Used by learners and educators for understanding and evaluating educational recommendations
Implementation: Developed as an interactive interface that allows for customizable explanations using AI and LLMs
Outcomes: High acceptance of tool components, user-friendliness, and increased motivation to explore recommendations
Challenges: Limitations in LLM's reliability and the need for further evidence-based experimentation
Implementation Barriers
Technical Limitations
LLMs may produce hallucinations or irrelevant responses in educational contexts, making them less reliable.
Proposed Solutions: Implementing contextual rules in prompt design and providing thorough support for LLM outputs to enhance reliability.
Project Team
Hasan Abu-Rasheed
Researcher
Christian Weber
Researcher
Madjid Fathi
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Hasan Abu-Rasheed, Christian Weber, Madjid Fathi
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