Disentangling Knowledge States with Ability and Proficiency Modeling for Knowledge Tracing

2026-07-14

AI Summary

This work addresses a critical limitation in existing knowledge tracing approaches, which model student interactions as a monolithic process and thereby overlook the distinct phases of competence acquisition and proficiency refinement, leading to biased representations of knowledge states. To remedy this, we propose Phase-Aware Knowledge Tracing (PAKT), the first framework to explicitly incorporate a dual-phase paradigm—separately modeling competence and proficiency. PAKT employs a tailored sequence decomposition mechanism to delineate learning phases and integrates a multi-branch Transformer architecture with a type-aware readout module to jointly capture phase-specific dynamics and holistic knowledge states. Through causal analysis, we identify and mitigate confounding bias arising from behavioral confounders in conventional models. Extensive experiments demonstrate that PAKT consistently outperforms state-of-the-art methods across six benchmark datasets, achieving average AUC gains of 0.82% and up to 1.33% in peak improvement.

Abstract

Knowledge tracing (KT) aims to predict students' future performance by modeling their evolving knowledge states from historical interactions. Existing KT methods usually treat the raw interaction sequence as a unified behavioral process, overlooking the phase-specific nature of learning behaviors. Our preliminary observations show that students are more likely to correctly answer previously failed knowledge concepts after sufficient practice, suggesting a transition from ability-building to proficiency-oriented learning. Motivated by this, we propose Phase-Aware Knowledge Tracing (PAKT), a KT framework that decomposes student interactions into ability and proficiency phases based on the tailored decomposition mechanism. To effectively exploit the decomposed sequences, we design a multi-branch Transformer with a type-aware readout module to jointly capture phase-specific and holistic knowledge states. We further provide a causal analysis to reveal the confounding bias caused by entangling complex learning behaviors in phase-agnostic KT models. Extensive experiments on six public benchmarks demonstrate that our method consistently outperforms representative baselines, with a maximum AUC gain of 1.33% and an average gain of 0.82%.

Problem

• Knowledge Tracing
• Learning Phases
• Ability Modeling
• Proficiency Modeling
• Knowledge States

Innovation

• Knowledge Tracing
• Phase-Aware Modeling
• Ability-Proficiency Decomposition
• Multi-branch Transformer
• Causal Analysis