Measuring Gender Bias in West Slavic Language Models
Publikation: Bidrag til bog/antologi/rapport › Konferencebidrag i proceedings › Forskning › fagfællebedømt
Dokumenter
- Fulltext
Forlagets udgivne version, 419 KB, PDF-dokument
Pre-trained language models have been known to perpetuate biases from the underlying datasets to downstream tasks. However, these findings are predominantly based on monolingual language models for English, whereas there are few investigative studies of biases encoded in language models for languages beyond English. In this paper, we fill this gap by analysing gender bias in West Slavic language models. We introduce the first template-based dataset in Czech, Polish, and Slovak for measuring gender bias towards male, female and non-binary subjects. We complete the sentences using both mono- and multilingual language models and assess their suitability for the masked language modelling objective. Next, we measure gender bias encoded in West Slavic language models by quantifying the toxicity and genderness of the generated words. We find that these language models produce hurtful completions that depend on the subject's gender. Perhaps surprisingly, Czech, Slovak, and Polish language models produce more hurtful completions with men as subjects, which, upon inspection, we find is due to completions being related to violence, death, and sickness.
Originalsprog | Engelsk |
---|---|
Titel | EACL 2023 - 9th Workshop on Slavic Natural Language Processing, Proceedings of the SlavicNLP 2023 |
Antal sider | 9 |
Forlag | Association for Computational Linguistics (ACL) |
Publikationsdato | 2023 |
Sider | 146-154 |
ISBN (Elektronisk) | 9781959429579 |
DOI | |
Status | Udgivet - 2023 |
Begivenhed | 9th Workshop on Slavic Natural Language Processing, SlavicNLP 2023 - Dubrovnik, Kroatien Varighed: 6 maj 2023 → … |
Konference
Konference | 9th Workshop on Slavic Natural Language Processing, SlavicNLP 2023 |
---|---|
Land | Kroatien |
By | Dubrovnik |
Periode | 06/05/2023 → … |
Bibliografisk note
Publisher Copyright:
© 2023 Association for Computational Linguistics.
ID: 372613271