Prerequisite chains are crucial to acquiring new knowledge efficiently. Many studies have been devoted to automatically identifying the prerequisite relationships between concepts from educational data. Though effective to some extent, these methods have neglected two key factors: most works have failed to utilize domain-related knowledge to enhance pre-trained language models, thus making the textual representation of concepts less effective; they also ignore the fusion of semantic information and structural information formed by existing prerequisites. We propose a two-stage concept prerequisite learning model (TCPL), to integrate the above factors. In the first stage, we designed two continual pre-training tasks for domain-adaptive and task-specific enhancement, to obtain better textual representation. In the second stage, to leverage the complementary effects of the semantic and structural information, we optimized the encoder of the resource–concept graph and the pre-trained language model simultaneously, with hinge loss as an auxiliary training objective. Extensive experiments conducted on three public datasets demonstrated the effectiveness of the proposed approach. Our proposed model improved by 7.9%, 6.7%, 5.6%, and 8.4% on ACC, F1, AP, and AUC on average, compared to the state-of-the-art methods.

Knowledge tracing (KT) refers to the problem of predicting a learner’s future performance based on their past performance in education. Recently, attention-based sequence modeling methods achieve impressive predictive performance. However, existing solutions merely consider one single sequence modeling method, which might fail to capture the comprehensive state of knowledge across long sequences. In this paper, we propose D ual S equence M odeling for K nowledge T racing (DSMKT). DSMKT aims to enhance the modeling of a learner’s long-term profile by collaborating two sequence modeling methods, i.e., the masked self-attention mechanism and the gated recurrent unit. To further exploit the synergy between two sequence models, we adopt the idea of online knowledge distillation and adaptively combine two branches to form a stronger teacher model, which in turn provides predictions as extra supervision for better modeling ability. Extensive experiments on four real-world benchmark datasets show that DSMKT performs excellently in predicting future learner responses.

Recommender systems that learn user preference from item-level feedback (provided to individual items) have been extensively studied. Considering the risk of privacy exposure, learning from set-level feedback (provided to sets of items) has been demonstrated to be a better way, since set-level feedback reveals user preferences while to some extent, hiding his/her privacy. Since only set-level feedback is provided as a supervision signal, different methods are being investigated to build connections between set-based preferences and item-based preferences. However, they overlook the complexity of user behavior in real-world applications. Instead, we observe that users’ set-level preference can be better modeled based on a subset of items in the original set. To this end, we propose to tackle the problem of identifying subsets from sets of items for set-based preference learning. We propose a policy network to explicitly learn a personalized subset selection strategy for users. Given the complex correlation between items in the set-rating process, we introduce a self-attention module to make sure all set members are considered in subset selecting process. Furthermore, we introduce gumble softmax to avoid gradient vanishing caused by binary selection in model learning. Finally the selected items are aggregated by user-specific personalized positional weights. Empirical evaluation with real-world datasets verifies the superiority of the proposed model over the state-of-the-art.