DiaCDM: Cognitive Diagnosis in Teacher-Student Dialogues using the Initiation-Response-Evaluation Framework

2025-09-29

AI Summary

Traditional cognitive diagnosis (CD) models struggle with the dynamic, unstructured nature of teacher-student dialogues and face challenges in semantic extraction. To address this, we propose the first dialogue-oriented cognitive diagnosis paradigm. Our method formalizes the IRE (Initiation–Response–Evaluation) pedagogical framework into a computable dialogue diagnostic structure and introduces a graph-based encoding scheme that explicitly models fine-grained associations among teacher questions, student responses, and domain concepts. This enables effective integration of multi-source semantic signals and yields interpretable knowledge-state representations. Experiments on three real-world teacher-student dialogue datasets demonstrate statistically significant improvements in diagnostic accuracy over state-of-the-art baselines. Moreover, our approach provides educators with transparent, traceable cognitive assessments—supporting actionable instructional decisions. The implementation is publicly available.

Technology Category

• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

While cognitive diagnosis (CD) effectively assesses students' knowledge mastery from structured test data, applying it to real-world teacher-student dialogues presents two fundamental challenges. Traditional CD models lack a suitable framework for handling dynamic, unstructured dialogues, and it's difficult to accurately extract diagnostic semantics from lengthy dialogues. To overcome these hurdles, we propose DiaCDM, an innovative model. We've adapted the initiation-response-evaluation (IRE) framework from educational theory to design a diagnostic framework tailored for dialogue. We also developed a unique graph-based encoding method that integrates teacher questions with relevant knowledge components to capture key information more precisely. To our knowledge, this is the first exploration of cognitive diagnosis in a dialogue setting. Experiments on three real-world dialogue datasets confirm that DiaCDM not only significantly improves diagnostic accuracy but also enhances the results' interpretability, providing teachers with a powerful tool for assessing students' cognitive states. The code is available at https://github.com/Mind-Lab-ECNU/DiaCDM/tree/main.

Problem

• Applying cognitive diagnosis to unstructured teacher-student dialogues
• Extracting diagnostic semantics accurately from lengthy educational dialogues
• Lacking suitable frameworks for dynamic dialogue-based cognitive assessment

Innovation

• Adapts IRE framework for dialogue cognitive diagnosis
• Develops graph-based encoding for teacher questions
• Integrates knowledge components to capture key information