Skip to main content Skip to navigation

Enhanced Classroom Dialogue Sequences Analysis with a Hybrid AI Agent: Merging Expert Rule-Base with Large Language Models

Project Overview

This document explores the integration of generative AI in education, focusing on a hybrid AI agent that merges expert rule-based systems with large language models (LLMs) to enhance classroom dialogue analysis. The study emphasizes the critical role of classroom dialogue in promoting student engagement and improving learning outcomes, while it also identifies the limitations of conventional dialogue analysis techniques. By accurately categorizing dialogue sequences, the hybrid AI agent not only bridges the gap between theoretical frameworks and practical classroom dynamics but also showcases high precision and reliability in its analyses. The findings suggest that this innovative approach can significantly benefit teaching practices, offering valuable support for teacher professional development and ultimately enhancing the educational experience for both educators and students.

Key Applications

Hybrid AI agent for classroom dialogue analysis

Context: Middle school classrooms, specifically in subjects like mathematics and Chinese lessons.

Implementation: The agent was developed by synthesizing expert rules from existing literature and integrating them with an LLM to analyze classroom dialogues.

Outcomes: High precision in coding dialogue sequences, significant time savings in analysis, and improved scalability for large datasets.

Challenges: Limitations in recognizing certain dialogue sequences and the generalisability of findings across diverse classroom contexts.

Implementation Barriers

Technical barrier

LLMs have not been extensively applied to recognize dialogue sequences, limiting their effectiveness in this area.

Proposed Solutions: Future research aims to bridge the gap between data-driven models and theoretical insights to enhance sequence recognition.

Methodological barrier

Existing automated methods often lack a comprehensive theoretical grounding, leading to challenges in aligning findings with pedagogical frameworks.

Proposed Solutions: Developing a hybrid approach that combines theoretical frameworks with data-driven insights.

Project Team

Yun Long

Researcher

Yu Zhang

Researcher

Contact Information

For information about the paper, please contact the authors.

Authors: Yun Long, Yu Zhang

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

Let us know you agree to cookies