Gender Bias in BERT - Measuring and Analysing Biases through Sentiment Rating in a Realistic Downstream Classification Task

Sophie Jentzsch, Cigdem Turan


Abstract
Pretrained language models are publicly available and constantly finetuned for various real-life applications. As they become capable of grasping complex contextual information, harmful biases are likely increasingly intertwined with those models. This paper analyses gender bias in BERT models with two main contributions: First, a novel bias measure is introduced, defining biases as the difference in sentiment valuation of female and male sample versions. Second, we comprehensively analyse BERT?s biases on the example of a realistic IMDB movie classifier. By systematically varying elements of the training pipeline, we can conclude regarding their impact on the final model bias. Seven different public BERT models in nine training conditions, i.e. 63 models in total, are compared. Almost all conditions yield significant gender biases. Results indicate that reflected biases stem from public BERT models rather than task-specific data, emphasising the weight of responsible usage.
Anthology ID:
2022.gebnlp-1.20
Volume:
Proceedings of the 4th Workshop on Gender Bias in Natural Language Processing (GeBNLP)
Month:
July
Year:
2022
Address:
Seattle, Washington
Venue:
GeBNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
184–199
Language:
URL:
https://aclanthology.org/2022.gebnlp-1.20
DOI:
10.18653/v1/2022.gebnlp-1.20
Bibkey:
Cite (ACL):
Sophie Jentzsch and Cigdem Turan. 2022. Gender Bias in BERT - Measuring and Analysing Biases through Sentiment Rating in a Realistic Downstream Classification Task. In Proceedings of the 4th Workshop on Gender Bias in Natural Language Processing (GeBNLP), pages 184–199, Seattle, Washington. Association for Computational Linguistics.
Cite (Informal):
Gender Bias in BERT - Measuring and Analysing Biases through Sentiment Rating in a Realistic Downstream Classification Task (Jentzsch & Turan, GeBNLP 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.gebnlp-1.20.pdf
Note:
 2022.gebnlp-1.20.note.pdf
Data
IMDb Movie Reviews