In e-learning, the increasing number of learning resources makes it difficult for learners to find suitable resources. addition, may have different preferences and cognitive abilities resources, where differences in learners’ will lead importance Therefore, recommending personalized paths has become a research hotspot. Considering this paper proposes path recommendation algorithm based on knowledge graph. We construct multi-dimensional courses graph computer field (MCCKG), then propose...

The multimodal knowledge graph link prediction model integrates entity features from multiple modalities, such as text and images, uses these fused to infer potential links in the graph. This process is highly dependent on fitting generalization capabilities of deep learning models, enabling models accurately capture complex semantic relational patterns. However, it this reliance that leads black-box nature decision-making mechanisms bases within which are difficult understand intuitively....

The knowledge graph of computer discipline domain plays a critical role in education, and the person event is an important part graph. Adding events to will make richer more interesting, enhance enthusiasm students for learning. most crucial step building extraction trigger words. Therefore, this paper proposes method based on serial fusion gated recurrent neural network convolutional (SC-BiGRU-CNN) detection domain. We extract global features text from sentences through BiGRU model, input...

Multimodal entity linking aims to link mentions target entities in the multimodal knowledge graph. The current mainly focuses on global fusion of text and image, seldom fully exploring correlation between modalities. In order improve effect feature, we propose a model based Co-Attention Fusion strategy. This strategy is designed enable image guide each other for extracting features, thus making full exploration modalities fine-grained feature effect. Furthermore, also design candidate...

Entity alignment is an important task in knowledge fusion, which aims to link entities that have the same real-world identity two graphs. However, process of constructing a graph, some noise may inevitably be introduced, must affect results entity tasks. The triple confidence calculation can quantify correctness triples reduce impact on alignment. Therefore, we designed method calculate and applied it representation learning phase calculates based pairing rates three angles between...

Entities refer to things that exist objectively, and entity types are concepts abstracted from entities have the same features or properties. However, in knowledge graph always incomplete. Currently, main approach for predicting missing is learn structured representations of separately, which ignores neighborhood semantic entity. Therefore, this paper proposes aggregation semantics model type completion (ANSTC), extracts triple target with two attentional mechanisms. Meanwhile, spatial...