Given the growing concern about the negative impact that gender-biased language embeddings may have in practical applications, we measure the gender bias in existing Romanian language embeddings using the method proposed in [1] for languages with grammatical gender, and invite contributors to submit debiasing methods that can lower the gender bias in existing embeddings, or submit less biased embeddings.
- The experiment was adapted for Romanian language from the original experiment showcased for Spanish and French.
- All the credits for adapting and running the experiments goes to Beata Lorincz.
1 Pei Zhou, Weijia Shi, Jieyu Zhao, Kuan-Hao Huang, Muhao Chen, Ryan Cotterell, and Kai-Wei Chang. 2019. Examining gender bias in languages with grammatical gender. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pages 5276–5284, Hong Kong, China. Association for Computational Linguistics.↩