References & Citations
Computer Science > Computation and Language
Title: From Alignment to Assignment: Frustratingly Simple Unsupervised Entity Alignment
(Submitted on 6 Sep 2021 (v1), last revised 15 Sep 2021 (this version, v2))
Abstract: Cross-lingual entity alignment (EA) aims to find the equivalent entities between crosslingual KGs, which is a crucial step for integrating KGs. Recently, many GNN-based EA methods are proposed and show decent performance improvements on several public datasets. Meanwhile, existing GNN-based EA methods inevitably inherit poor interpretability and low efficiency from neural networks. Motivated by the isomorphic assumption of GNNbased methods, we successfully transform the cross-lingual EA problem into the assignment problem. Based on this finding, we propose a frustratingly Simple but Effective Unsupervised entity alignment method (SEU) without neural networks. Extensive experiments show that our proposed unsupervised method even beats advanced supervised methods across all public datasets and has high efficiency, interpretability, and stability.
Submission history
From: Xin Mao [view email][v1] Mon, 6 Sep 2021 11:02:26 GMT (480kb,D)
[v2] Wed, 15 Sep 2021 03:24:33 GMT (463kb,D)
Link back to: arXiv, form interface, contact.