Unsupervised Entity Alignment Based on Personalized Discriminative Rooted Tree

2025-02-14

AI Summary

This work addresses two key challenges in unsupervised entity alignment (EA): (1) embedding homogenization, caused by shared aggregation paths in graph neural networks (GNNs) that hinder entity-level personalization; and (2) cross-graph distribution shift, exacerbated by the absence of supervision and leading to structural mismatch between knowledge graphs. To tackle these, we propose a personalized root-tree modeling framework: discriminative root-tree sampling constructs entity-specific neighborhoods, and tree-based attention aggregation generates distinctive embeddings; additionally, a mutual information maximization (MIM) auxiliary task explicitly enforces cross-graph embedding distribution consistency in a fully unsupervised manner. Our approach eliminates reliance on seed alignments and establishes the first EA paradigm integrating personalized topological modeling with unsupervised mutual information regularization. Experiments demonstrate new state-of-the-art performance for unsupervised EA across multiple benchmarks, surpassing most supervised baselines.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Reasoning under Uncertainty: Graphical Models

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Entity Alignment (EA) is to link potential equivalent entities across different knowledge graphs (KGs). Most existing EA methods are supervised as they require the supervision of seed alignments, i.e., manually specified aligned entity pairs. Very recently, several EA studies have made some attempts to get rid of seed alignments. Despite achieving preliminary progress, they still suffer two limitations: (1) The entity embeddings produced by their GNN-like encoders lack personalization since some of the aggregation subpaths are shared between different entities. (2) They cannot fully alleviate the distribution distortion issue between candidate KGs due to the absence of the supervised signal. In this work, we propose a novel unsupervised entity alignment approach called UNEA to address the above two issues. First, we parametrically sample a tree neighborhood rooted at each entity, and accordingly develop a tree attention aggregation mechanism to extract a personalized embedding for each entity. Second, we introduce an auxiliary task of maximizing the mutual information between the input and the output of the KG encoder, to regularize the model and prevent the distribution distortion. Extensive experiments show that our UNEA achieves a new state-of-the-art for the unsupervised EA task, and can even outperform many existing supervised EA baselines.

Problem

• Unsupervised entity alignment in knowledge graphs
• Personalized entity embeddings extraction
• Mitigation of distribution distortion between KGs

Innovation

• Unsupervised entity alignment technique
• Personalized tree attention aggregation
• Mutual information maximization regularization