Raoufi E. auteur aut IRD Happi Happi Bill Gates auteur aut IRD Larmande Pierre auteur aut IRD Scharffe F. auteur aut IRD Todorov K. auteur aut IRD text journalArticle eng text/pdf reformatted digital access Representation learning for entity alignment (EA) aims to identify entities in different knowledge graphs (KGs) that refer to the same real-world object by comparing their embedding similarity. Although many EA models perform well on synthetic benchmark datasets, this performance does not always transfer to real-world, incomplete, and domain-specific data. A systematic comparison between synthetic benchmarks and original heterogeneous datasets is still limited. Many EA models also restrict the alignment search space to validation entities, limiting coverage of real KG content. Within this setting, our results show that embedding-based EA models continue to face generalization challenges in realistic large-scale KG search spaces. We evaluate several competitive EA models-commonly tested on benchmarks such as DBP15K-on multiple real-world heterogeneous datasets. The experiments reveal a performance decrease when moving beyond synthetic benchmarks, indicating that current models do not fully capture the characteristics of real data. We also analyze semantic similarity and profiling features of the datasets to help explain these differences. This study outlines practical limitations of embedding-based EA methods and provides insights for developing approaches that better handle the variability and complexity found in real-world KG alignment tasks. specialized entity alignment knowledge graphs representation learning knowledge graph heterogeneity EA benchmarks 122 020 17 1 09217134251389825 [24 ] 2026 1570-0844 https://www.documentation.ird.fr/hor/fdi:010095821 10.1177/09217134251389825 1570-0844 [F B010095821] https://www.documentation.ird.fr/hor/fdi:010095821 https://horizon.documentation.ird.fr/exl-doc/pleins_textes/2026-01/010095821.pdf IRD - Base Horizon / Pleins textes 2026-01-14 2026-01-19 fdi:010095821 fre