A path-based relation networks model for knowledge graph completion

Lee, Wan-Kon; Shin, Won-Chul; Jagvaral, Batselem; Roh, Jae-Seung; Kim, Min-Sung; Park, Hyunkyu; Park, Young-Tack

doi:10.1016/j.eswa.2021.115273

Cited by 14 publications

(2 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Nathani et al 16 proposed KBGAT, which defines an attention layer and uses a hierarchical iteration to spread attention based on the concatenation of entity and relation embeddings. Lee et al 17 proposed a link prediction method based on relation paths. This method transforms the graph structure into path sequences using a path ranking algorithm and predicts missing entities and relations through sequence models.…”

Section: Kgc Based On Graph Embeddingmentioning

confidence: 99%

A novel model for relation prediction in knowledge graphs exploiting semantic and structural feature integration

Yang,

Lu,

et al. 2024

Sci Rep

View full text Add to dashboard Cite

Relation prediction is a critical task in knowledge graph completion and associated downstream tasks that rely on knowledge representation. Previous studies indicate that both structural features and semantic information are meaningful for predicting missing relations in knowledge graphs. This has led to the development of two types of methods: structure-based methods and semantics-based methods. Since these two approaches represent two distinct learning paradigms, it is difficult to fully utilize both sets of features within a single learning model, especially deep features. As a result, existing studies usually focus on only one type of feature. This leads to an insufficient representation of knowledge in current methods and makes them prone to overlooking certain patterns when predicting missing relations. In this study, we introduce a novel model, RP-ISS, which combines deep semantic and structural features for relation prediction. The RP-ISS model utilizes a two-part architecture, with the first component being a RoBERTa module that is responsible for extracting semantic features from entity nodes. The second part of the system employs an edge-based relational message-passing network designed to capture and interpret structural information within the data. To alleviate the computational burden of the message-passing network on the RoBERTa module during the sampling process, RP-ISS introduces a node embedding memory bank, which updates asynchronously to circumvent excessive computation. The model was assessed on three publicly accessible datasets (WN18RR, WN18, and FB15k-237), and the results revealed that RP-ISS surpasses all baseline methods across all evaluation metrics. Moreover, RP-ISS showcases robust performance in graph inductive learning.

show abstract

Section: Kgc Based On Graph Embeddingmentioning

confidence: 99%

A novel model for relation prediction in knowledge graphs exploiting semantic and structural feature integration

Yang,

Lu,

et al. 2024

Sci Rep

View full text Add to dashboard Cite

show abstract

“…In this structure the CBOW and CNN connectors take blank texts like embedding and business embedding as output to minimize imitation of the homework mentioned above. In order to maximize efficiency, we also use CNN's multimedia version to read presentations (Lee et al, 2021). The general backward rotation using a stochastic gradient descent (SGD) with the Rear propagation will be prevented if it meets a zero finish or the current value of the feature was not considered in the compilation during the forward broadcast (Lee et al, 2021).…”

Section: Initialization Of the Descriptive Modelmentioning

confidence: 99%

Learning of Knowledge Graphs with Entity Descriptions

Conde¹

2022

IJMRA

View full text Add to dashboard Cite

The representing learning makes specialty of knowledge graph and it indicates the difference between different entities. The knowledge graph representing with the low-dimensional space. In fact, most of the method usually concern of description of entity which is hard for existing strategies to take benefit of. Here, we recommend a new representing learning method with knowledge graphs that uses entity description. We evaluate our method on two tasks like knowledge graph and entity classification. Experimental effects on actual-world datasets show that our version plays higher than different baseline fashions, especially under the zero-short setting, which indicate that our technique for novel the entity description.

show abstract