UNVaMP: Neural Knowledge Tracing with Variational Regularization of Latent Knowledge Dynamics

2026-08-04

AI Summary

This work addresses the challenge in knowledge tracing of simultaneously achieving high prediction accuracy, interpretability, smooth learning trajectories, and principled uncertainty quantification. To this end, the authors propose UNVaMP, a unified neural variational framework that integrates student-item interactions with internal memory to construct dynamically evolving latent knowledge representations through variational regularization. This approach enables both response prediction and controllable smoothing of knowledge trajectories. The model supports both purely neural and hybrid interpretable configurations—such as UNVaMP-MIRT—and achieves state-of-the-art prediction performance on three out of four real-world datasets. While the hybrid variant slightly underperforms the purely neural version in predictive accuracy, it yields interpretable estimates of knowledge states and effectively quantifies uncertainty. Empirical results further demonstrate the model’s sensitivity to structural information and its robustness.

Abstract

We introduce the Unified Neural Variational Measurement of Proficiency (UNVaMP) architecture, a knowledge tracing method that integrates observed student-item interactions with internal memory to produce evolving latent representations of student knowledge. These representations support accurate predictions of future responses while enabling explicit control over the smoothness of estimated learning trajectories. UNVaMP can be configured as either a purely neural model or a hybrid model that predicts responses through an interpretable measurement function over the latent space. We show that a pure neural configuration (UNVaMP-MLP) achieves the strongest predictive performance among compared models on three out of four datasets. Meanwhile, a hybrid configuration (UNVaMP-MIRT, using a 1PL MIRT measurement function) lags only slightly behind UNVaMP-MLP, indicating that the predictive cost of interpretability is modest. Beyond predictive accuracy, UNVaMP provides the following: a principled mechanism for controlling volatility when estimating student latent variables, quantification of uncertainty over student knowledge state estimates, and flexible input specification that supports heterogeneous student-item interaction features. In addition, the hybrid UNVaMP-MIRT configuration generates interpretable moment-in-time student knowledge state estimates. Using an experimental dataset, we show that auxiliary inputs induce structured changes in the predictive behavior of UNVaMP-MIRT, consistent with sensitivity to underlying structure beyond response correctness. Furthermore, through a simulation study, we show that UNVaMP yields well-behaved knowledge state estimates under controlled measurement conditions. In total, these results indicate that UNVaMP is both useful for real-world education systems and capable of recovering underlying structure from student-item interactions.

Problem

• knowledge tracing
• latent knowledge dynamics
• predictive accuracy
• interpretability
• uncertainty quantification

Innovation

• knowledge tracing
• variational regularization
• latent knowledge dynamics
• interpretable modeling
• neural measurement