.png){kind=icon}
.png){kind=icon}
Return
Hyperbolic graph representation learning: methods, applications and challenges—A survey
.png){kind=icon}
PRE
AI
DOI：10.1016/j.neucom.2025.131044.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Graph-structured data is widely used in real-world applications such as social networks, recommender systems, and transportation networks. With the rapid expansion of network scale and the continuous emergence of massive data, graph representation learning has become a research hotspot in the field of graph data mining. Current research on graph representation learning focuses on graph features in Euclidean space. These investigations are simple and efficient but hard to accurately capture the potential power-law distribution and hierarchical structure characteristics of complex networks, which fail to achieve lower distortion when embedding tree-like structured networks. The exponential growth property of hyperbolic geometry is suitable for representing tree-like hierarchical networks and provides unique advantages in embedding heterogeneous or scale-free graphs, making hyperbolic space a novel embedding space for graph representation learning. In this survey, we present a comprehensive overview of the geometric knowledge, methods, tools, and applications on Hyperbolic Graph Representation Learning (HGRL). Firstly, the geometric knowledge involved in HGRL is summarized, with an emphasis on the hyperbolic models. Secondly, we propose a new taxonomy of methods for HGRL and illustrate some examples. Next, we introduce two tools applied in HGRL. Then, the current applications of HGRL in different fields are examined. Finally, directions for further research are outlined, providing insights into future advancements in HGRL.
Keywords:
graph representation learning
hyperbolic geometry
hierarchical networks
graph embedding
scale-free graphs