Generative Cognitive Diagnosis

2025-07-13

AI Summary

Traditional cognitive diagnosis models (CDMs) require costly retraining for new learners, resulting in high computational overhead and suboptimal diagnostic reliability. To address this, we propose a generative paradigm for cognitive diagnosis—shifting from discriminative prediction modeling to generative modeling for the first time. Our approach decouples cognitive state inference from response generation, enabling instantaneous, parameter-free diagnosis. Methodologically, we introduce Generative Item Response Theory (G-IRT) and a Generative Neural Cognitive Diagnosis Model (G-NCDM), incorporating identifiability and monotonicity constraints to ensure statistical validity of the generative process. Evaluated on real-world educational datasets, our models achieve up to 100× faster diagnosis while significantly outperforming baseline CDMs in both accuracy and stability. This work establishes a novel, efficient, reliable, and scalable paradigm for intelligent educational systems and AI model evaluation.

Technology Category

• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Machine Learning: Deep Generative Models & Autoencoders

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

Cognitive diagnosis (CD) models latent cognitive states of human learners by analyzing their response patterns on diagnostic tests, serving as a crucial machine learning technique for educational assessment and evaluation. Traditional cognitive diagnosis models typically follow a transductive prediction paradigm that optimizes parameters to fit response scores and extract learner abilities. These approaches face significant limitations as they cannot perform instant diagnosis for new learners without computationally expensive retraining and produce diagnostic outputs with limited reliability. In this study, we introduces a novel generative diagnosis paradigm that fundamentally shifts CD from predictive to generative modeling, enabling inductive inference of cognitive states without parameter re-optimization. We propose two simple yet effective instantiations of this paradigm: Generative Item Response Theory (G-IRT) and Generative Neural Cognitive Diagnosis Model (G-NCDM), which achieve excellent performance improvements over traditional methods. The generative approach disentangles cognitive state inference from response prediction through a well-designed generation process that incorporates identifiability and monotonicity conditions. Extensive experiments on real-world datasets demonstrate the effectiveness of our methodology in addressing scalability and reliability challenges, especially $ imes 100$ speedup for the diagnosis of new learners. Our framework opens new avenues for cognitive diagnosis applications in artificial intelligence, particularly for intelligent model evaluation and intelligent education systems. The code is available at https://github.com/CSLiJT/Generative-CD.git.

Problem

• Enables instant diagnosis for new learners without retraining
• Improves reliability and scalability of cognitive state inference
• Shifts cognitive diagnosis from predictive to generative modeling

Innovation

• Generative paradigm shifts CD to inductive inference
• G-IRT and G-NCDM improve performance significantly
• Well-designed generation ensures identifiability and monotonicity