EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-SA 4.0

Name-Based Embedding Debiasing: Analyzing Its Impact on Gender, Religious, and Ethnic Biases

Gianmarco Cafferata, Mariano G. Beiró

SADIO Electronic Journal of Informatics and Operations Research · 27.05.2026 · 25 (2) · e102

DOI: 10.24215/15146774e102 ↗

Kurz gefasst

Fachartikel in der Zeitschrift SADIO Electronic Journal of Informatics and Operations Research, online erschienen am 27.05.2026. Verfasst von Gianmarco Cafferata und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-SA 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Word vector representations were the initial building block that started the current state-of-the-art methods for several NLP tasks. Bias metrics and debiasing methods for static embeddings have been studied with moderate success, achieving some bias reductions for specific groups and metrics. However, these methods often fail to improve multiple metrics simultaneously or to meaningfully impact extrinsic tasks. Recent research in debiasing has mainly shifted its focus towards contextual embeddings and large language models (LLMs). Here we argue that static embeddings provide a simpler and more controlled setting for testing hypotheses and techniques, which can then be extended to more complex models. We introduce a method that captures multiple demographic dimensions (gender, race, age, etc.) in static embeddings simultaneously, eliminating the need for specialized tasks or demographic-specific vocabulary.

Zitieren

APA

Cafferata, G., & Beiró, M. G. (2026). Name-Based Embedding Debiasing: Analyzing Its Impact on Gender, Religious, and Ethnic Biases. SADIO Electronic Journal of Informatics and Operations Research, 25(2), e102. https://doi.org/10.24215/15146774e102

Vancouver

Cafferata G, Beiró MG. Name-Based Embedding Debiasing: Analyzing Its Impact on Gender, Religious, and Ethnic Biases. SADIO Electronic Journal of Informatics and Operations Research. 2026;25(2):e102. doi:10.24215/15146774e102

BibTeX
@article{Cafferata2026,
  author = {Cafferata, Gianmarco and Beiró, Mariano G.},
  title = {Name-Based Embedding Debiasing: Analyzing Its Impact on Gender, Religious, and Ethnic Biases},
  journal = {SADIO Electronic Journal of Informatics and Operations Research},
  year = {2026},
  volume = {25},
  number = {2},
  pages = {e102},
  doi = {10.24215/15146774e102}
}
Kostenlos starten →