Towards Actionable Pedagogical Feedback: A Multi-Perspective Analysis of Mathematics Teaching and Tutoring Dialogue
Project Overview
The document examines the application of generative AI in mathematics education, specifically through the utilization of advanced natural language processing (NLP) models to analyze classroom dialogues. It underscores the significance of effective feedback in enhancing instructional practices and introduces a framework for multi-perspective discourse analysis aimed at improving teacher feedback and educational outcomes. The findings highlight the critical roles of both talk moves and non-talk moves within classroom interactions, demonstrating how these components can inform the development of AI-driven educational tools. By focusing on responsive learning environments, the study illustrates the potential of generative AI to transform educational experiences, ultimately contributing to more effective teaching strategies and improved student engagement in mathematics.
Key Applications
Multi-Perspective Discourse Analysis Framework
Context: K-12 mathematics teaching and tutoring
Implementation: Utilized two datasets (TalkMoves and SAGA22) to analyze teacher and student dialogues, employing NLP models to classify talk moves, dialogue acts, and discourse relations.
Outcomes: Improved understanding of classroom discourse dynamics and enhanced feedback mechanisms for educators; identified meaningful interaction patterns to inform instructional strategies.
Challenges: Multi-functionality of utterances complicating classification; non-targeted moves often ignored in analysis.
Implementation Barriers
Technical Barrier
Multi-functionality of utterances where a single utterance may serve multiple purposes, complicating analysis.
Proposed Solutions: Designing multiple indicators to capture various functions of utterances.
Data Barrier
Many non-targeted utterances are excluded from discourse analysis, leading to gaps in feedback.
Proposed Solutions: Incorporating comprehensive analysis that includes non-targeted moves to improve the feedback provided.
Project Team
Jannatun Naim
Researcher
Jie Cao
Researcher
Fareen Tasneem
Researcher
Jennifer Jacobs
Researcher
Brent Milne
Researcher
James Martin
Researcher
Tamara Sumner
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Jannatun Naim, Jie Cao, Fareen Tasneem, Jennifer Jacobs, Brent Milne, James Martin, Tamara Sumner
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