Aligning Multiple Knowledge Graphs in a Single Pass

TL;DR

The paper tackles the problem of aligning more than two knowledge graphs by introducing MultiEA, a framework that embeds entities from all candidate KGs into a shared embedding space using a single, shared KG encoder. It explores three alignment strategies to minimize distances among pre-aligned entities and adds an inference enhancement that incorporates higher-order similarities to boost performance. Two real-world benchmarks, DBP-4 and DWY-3, demonstrate that MultiEA achieves effective and efficient multi-KG alignment in a single pass, with substantial improvements over baselines and favorable efficiency trade-offs. The authors provide open-source code to advance research in multi-KG alignment and cross-KG fusion.

Abstract

Entity alignment (EA) is to identify equivalent entities across different knowledge graphs (KGs), which can help fuse these KGs into a more comprehensive one. Previous EA methods mainly focus on aligning a pair of KGs, and to the best of our knowledge, no existing EA method considers aligning multiple (more than two) KGs. To fill this research gap, in this work, we study a novel problem of aligning multiple KGs and propose an effective framework named MultiEA to solve the problem. First, we embed the entities of all the candidate KGs into a common feature space by a shared KG encoder. Then, we explore three alignment strategies to minimize the distances among pre-aligned entities. In particular, we propose an innovative inference enhancement technique to improve the alignment performance by incorporating high-order similarities. Finally, to verify the effectiveness of MultiEA, we construct two new real-world benchmark datasets and conduct extensive experiments on them. The results show that our MultiEA can effectively and efficiently align multiple KGs in a single pass. We release the source codes of MultiEA at: https://github.com/kepsail/MultiEA.

Figures (6)

• Figure 1: An example of aligning four KGs.
• Figure 2: (a) Moving toward the mean (the central grey square with dashed line); (b) Moving toward the anchor (the central blue square); (c) Moving toward each other.
• Figure 3: The entity names of the ten groups of equivalent entities discovered by MultiEA on DBP-4.
• Figure 4: The entity names of ten groups of equivalent entities discovered by MultiEA on DWY-3. Note that, for DWY-3: wiki, the original data is in the form of the URLs of the entities. We show the corresponding entries of the involved URLs.
• Figure 5: Visualization of the embeddings of the ten groups of equivalent entities listed in Fig. \ref{['fig:case-dbp']} and Fig. \ref{['fig:case-dwy']}.