Behavior-Aware Item Modeling via Dynamic Procedural Solution Representations for Knowledge Tracing

2026-04-09

AI Summary

This work addresses the limitation of existing knowledge tracing approaches in capturing procedural dynamics during problem solving, which hinders precise modeling of learners’ cognitive behaviors. To overcome this, the study introduces Pólya’s problem-solving framework into knowledge tracing for the first time, decomposing each solution into four stages: understanding, planning, execution, and review. Stage-level embedding trajectories are generated using a reasoning language model, and a context-aware adaptive fusion mechanism dynamically integrates representations from all stages to enhance behavior-aware modeling. Integrated with a pretrained knowledge tracing backbone, the proposed method significantly outperforms strong baselines on the XES3G5M and NIPS34 datasets, demonstrating particularly notable gains in scenarios involving repeated learner-item interactions.

Abstract

Knowledge Tracing (KT) aims to predict learners' future performance from past interactions. While recent KT approaches have improved via learning item representations aligned with Knowledge Components, they overlook the procedural dynamics of problem solving. We propose Behavior-Aware Item Modeling (BAIM), a framework that enriches item representations by integrating dynamic procedural solution information. BAIM leverages a reasoning language model to decompose each item's solution into four problem-solving stages (i.e., understand, plan, carry out, and look back), pedagogically grounded in Polya's framework. Specifically, it derives stage-level representations from per-stage embedding trajectories, capturing latent signals beyond surface features. To reflect learner heterogeneity, BAIM adaptively routes these stage-wise representations, introducing a context-conditioned mechanism within a KT backbone, allowing different procedural stages to be emphasized for different learners. Experiments on XES3G5M and NIPS34 show that BAIM consistently outperforms strong pretraining-based baselines, achieving particularly large gains under repeated learner interactions.

Problem

• Knowledge Tracing
• Item Modeling
• Procedural Dynamics
• Problem Solving Stages
• Learner Heterogeneity

Innovation

• Knowledge Tracing
• Procedural Solution Representation
• Behavior-Aware Modeling
• Polya's Problem-Solving Framework
• Adaptive Stage Routing