Abstract:The research on knowledge graph completion based on representation learning is increasingly dependent on the node structural feature in the graph. However, a large number of nodes have few immediate neighbors, resulting in the node features unable to be fully expressed. Hence, multi-hop structure features are crucial to the representation learning of nodes. GCN (Graph Convolutional Network) is a graph embedding model that can introduce the multi-hop structure. However, the multi-hop information transmitted between GCN layers suffers a lot of losses. This would lead to the insufficient mining of the node structure features and semantic feature association among entities, further reducing the efficiency of graph knowledge completion. A gate-controlled graph sub-hop convolutional network model for knowledge graph completion is proposed to fill these research gaps. Firstly, a graph sub-hop convolutional network based on matrix representation is designed, which can transmit multi-hop neighbor features directly to the encoded node vector to avoid a large loss of features during multi-hop transmission. On this basis, the implicit multi-hop relations are explicitly embedded into the model based on the TransE. In the process of each hop convolution, aiming at the accumulation of noise redundancy caused by the increase of the receptive field, a sub-hop gate mechanism strategy is proposed to filter information. Finally, the linear model is used to decode the encoded nodes and then complete the knowledge graph. We carried out experimental comparison and analysis on WN18RR, FB15k-237, UMLS, and KINSHIP datasets. The results show that the embedding method based on the sub-hop structural information fusion can greatly improve the results of link prediction.

Exploiting Node-Feature Bipartite Graph in Graph Convolutional Networks.

Joint Learning of Graph Representation and Node Features in Graph Convolutional Neural Networks

Bipartite Graph Neural Networks for Efficient Node Representation Learning

A Graph Convolutional Neural Network Model Based on Fused Multi-Subgraph As Input and Fused Feature Information As Output

Dual Feature Interaction-Based Graph Convolutional Network

A simple and effective convolutional operator for node classification without features by graph convolutional networks

Unified Structure-Aware Feature Learning for Graph Convolutional Network

CoNet: Co-Embedding by Reinforcing Graph Feature and Topology Information

Feature Expansion for Graph Neural Networks

Exploiting Edge Features in Graph Neural Networks

An Embedding Model for Knowledge Graph Completion Based on Graph Sub-Hop Convolutional Network

Bilinear Graph Neural Network with Neighbor Interactions

Neighborhood Convolutional Graph Neural Network

Exploiting Edge Features for Graph Neural Networks

WGCN: Graph Convolutional Networks with Weighted Structural Features

Convolution Based Graph Representation Learning from the Perspective of High Order Node Similarities

Cross-GCN: Enhancing Graph Convolutional Network with K-Order Feature Interactions

Joint graph and feature learning in graph convolutional neural networks

Graph Neural Networks Provably Benefit from Structural Information: A Feature Learning Perspective

Self-Attention Empowered Graph Convolutional Network for Structure Learning and Node Embedding

Neighborhood Convolutional Network: A New Paradigm of Graph Neural Networks for Node Classification