Knowledge Graphs and Graph Neural Networks: A Detailed Survey on GNN Applications

Graph Neural Networks (GNNs) have become a significant focus in the field of Knowledge Graphs (KGs) due to their effectiveness in modeling graph-structured data. Published on May 12, 2026, this comprehensive survey by Chengcheng Sun and colleagues addresses the gaps in existing literature regarding GNN-based methodologies throughout the knowledge graph technologies pipeline.

Understanding the Knowledge Graph Technologies Pipeline

The survey introduces a novel two-level taxonomy framework for GNN-based knowledge graph technologies. This framework encompasses the full knowledge graph technologies pipeline, which includes:

• Knowledge graph construction
• Knowledge graph embedding
• Knowledge reasoning
• Knowledge graph applications

By outlining these stages, the authors provide a clear understanding of how GNNs can enhance various aspects of knowledge graphs.

GNN Models in Knowledge Graph Technologies

The paper categorizes GNN-based models such as GCN (Graph Convolutional Network), GAT (Graph Attention Network), and HGNN (Heterogeneous Graph Neural Network). Each model is analyzed concerning its strengths and limitations, presenting a detailed review of their applications across different tasks in the knowledge graph lifecycle.

Furthermore, the authors discuss the advantages of utilizing GNN technology based on the unique characteristics of various tasks, offering insights into how these models can be effectively implemented.

Future Directions and Challenges in GNN-Based Knowledge Graphs

Despite the advancements, the authors highlight several unresolved challenges within the field of GNN-based knowledge graphs. They outline promising directions for future research, encouraging further exploration into the potential of GNNs to revolutionize knowledge graph applications.

This survey, accepted for publication in ACM Computing Surveys, serves as a critical resource for researchers and practitioners interested in the intersection of GNNs and knowledge graphs.