Enhancing Knowledge Tracing via Adversarial Training

Abstract

We study the problem of knowledge tracing (KT) where the goal is to trace the students' knowledge mastery over time so as to make predictions on their future performance. Owing to the good representation capacity of deep neural networks (DNNs), recent advances on KT have increasingly concentrated on exploring DNNs to improve the performance of KT. However, we empirically reveal that the DNNs based KT models may run the risk of overfitting, especially on small datasets, leading to limited generalization. In this paper, by leveraging the current advances in adversarial training (AT), we propose an efficient AT based KT method (ATKT) to enhance KT model's generalization and thus push the limit of KT. Specifically, we first construct adversarial perturbations and add them on the original interaction embeddings as adversarial examples. The original and adversarial examples are further used to jointly train the KT model, forcing it is not only to be robust to the adversarial examples, but also to enhance the generalization over the original ones. To better implement AT, we then present an efficient attentive-LSTM model as KT backbone, where the key is a proposed knowledge hidden state attention module that adaptively aggregates information from previous knowledge hidden states while simultaneously highlighting the importance of current knowledge hidden state to make a more accurate prediction. Extensive experiments on four public benchmark datasets demonstrate that our ATKT achieves new state-of-the-art performance. Code is available at: https://github.com/xiaopengguo/ATKT. CCS CONCEPTS• Applied computing → Computer-assisted instruction; Learning management systems.

pyKT: A Python Library to Benchmark Deep Learning based Knowledge Tracing Models

“…• Deep sequential models: the chronologically ordered interaction sequence is captured by deep sequential models such as LSTM and GRU [4,11,14,15,20,21,25,34,45]. Selected approaches are DKT [25], DKT+ [45], DKT-F [21], and KQN [14].…”

“…Selected approach is DKVMN [46]. • Adversarial based models: the adversarial training techniques such as adversarial perturbations are applied into the original student interaction sequence to reduce the the risk of DLKT overfitting and limited generalization problem [11]. Selected approach is ATKT [11].…”

“…• Adversarial based models: the adversarial training techniques such as adversarial perturbations are applied into the original student interaction sequence to reduce the the risk of DLKT overfitting and limited generalization problem [11]. Selected approach is ATKT [11]. • Graph based models: the response interactions between students and questions and the knowledge associations between questions and KCs form a tripartite graph and graph based techniques are applied to aggregate such relations [22,37,44].…”

“…Similar to all existing DLKT research works [1,4,10,11,14,15,20,21,22,24,25,26,29,34,37,44,45,46,47], we use the area under the receiver operating characteristics curve (AUC) as the main metric to evaluate the performance of DLKT models on predicting binary-valued future learner responses to either questions or KCs. Meanwhile, we also include detailed model performance in terms of accuracy in Appendix.…”

“…Since then, many attempts have been made to solve the KT problem, such as probabilistic graphical models [12] and factor analysis based models [3,13,35]. Recently, due to the rapid advances of deep neural networks, deep learning based knowledge tracing (DLKT) models have become the de facto KT framework for modeling students' mastery of KCs [1,7,10,11,14,15,18,21,22,23,24,25,29,30,31,36,37,44,45,46,47].…”