Abstract: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.

BERT-enhanced Relational Sentence Ordering Network.

Efficient Relational Sentence Ordering Network

Deep Attentive Sentence Ordering Network.

Enriching Pre-trained Language Model with Entity Information for Relation Classification

End-to-End Neural Sentence Ordering Using Pointer Network.

Siamese BERT Model with Adversarial Training for Relation Classification.

Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks

BART-based Hierarchical Attentional Network for Sentence Ordering

Non-Autoregressive Sentence Ordering

Coreferential Reasoning Learning for Language Representation

Local and Global Context-Based Pairwise Models for Sentence Ordering

Exploring Effective Inter-Encoder Semantic Interaction for Document-Level Relation Extraction

RpBERT: A Text-image Relation Propagation-based BERT Model for Multimodal NER

CoT-BERT: Enhancing Unsupervised Sentence Representation through Chain-of-Thought

Enhancing Word-Level Semantic Representation via Dependency Structure for Expressive Text-to-Speech Synthesis

RelBERT: Embedding Relations with Language Models

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

SBERT-WK: A Sentence Embedding Method by Dissecting BERT-based Word Models

BART-Reader: Predicting Relations Between Entities via Reading Their Document-Level Context Information

PromptBERT: Improving BERT Sentence Embeddings with Prompts

A neuralized feature engineering method for entity relation extraction