Unifying Dual-Space Embedding for Entity Alignment via Contrastive Learning
Cunda Wang, Weihua Wang, Qiuyu Liang, Feilong Bao, Guanglai Gao
TL;DR
This work tackles entity alignment across knowledge graphs by unifying dual-space embeddings in Euclidean and hyperbolic spaces. It introduces UniEA, which jointly learns Euclidean neighborhood information with GATs and hierarchical structure in hyperbolic space via HGCNs, while enforcing cross-space consistency through contrastive losses. A margin-based alignment loss using seed pairs and negative sampling further strengthens cross-graph matching, and an intra-graph contrastive term discourages excessive similarity among neighboring entities. Extensive experiments on OpenEA datasets demonstrate state-of-the-art performance and robust ablations confirm the contribution of dual-space learning and contrastive objectives. The approach offers a scalable, structure-preserving pathway for cross-graph entity alignment with practical impact on knowledge integration tasks.
Abstract
Entity alignment aims to match identical entities across different knowledge graphs (KGs). Graph neural network-based entity alignment methods have achieved promising results in Euclidean space. However, KGs often contain complex structures, including both local and hierarchical ones, which make it challenging to efficiently represent them within a single space. In this paper, we proposed a novel method UniEA, which unifies dual-space embedding to preserve the intrinsic structure of KGs. Specifically, we learn graph structure embedding in both Euclidean and hyperbolic spaces simultaneously to maximize the consistency between the embedding in both spaces. Moreover, we employ contrastive learning to mitigate the misalignment issues caused by similar entities, where embedding of similar neighboring entities within the KG become too close in distance. Extensive experiments on benchmark datasets demonstrate that our method achieves state-of-the-art performance in structure-based EA. Our code is available at https://github.com/wonderCS1213/UniEA.