Iterative Adversarial Removal of Gender Bias in Pretrained Word Embeddings - Archive ouverte HAL Access content directly
Conference Papers Year : 2022

Iterative Adversarial Removal of Gender Bias in Pretrained Word Embeddings

Abstract

Recent advances in Representation Learning have discovered a strong inclination for pre-trained word embeddings to demonstrate unfair and discriminatory gender stereotypes. These usually come in the shape of unjustified associations between representations of group words (e.g., male or female) and attribute words (e.g. driving, cooking, doctor, nurse, etc.) In this paper, we propose an iterative and adversarial procedure to reduce gender bias in word vectors. We aim to remove gender influence from word representations that should otherwise be free of it, while retaining meaningful gender information in words that are inherently charged with gender polarity (male or female). We confine these gender signals in a sub-vector of word embeddings to make them more interpretable. Quantitative and qualitative experiments confirm that our method successfully reduces gender bias in pre-trained word embeddings with minimal semantic offset.
Fichier principal
Vignette du fichier
SAC___Iterative_Adversarial_removal_of_gender_bias (1).pdf (834 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03626768 , version 1 (31-03-2022)

Identifiers

Cite

Yacine Gaci, Boualem Benatallah, Fabio Casati, Khalid Benabdeslem. Iterative Adversarial Removal of Gender Bias in Pretrained Word Embeddings. The 37th ACM/SIGAPP Symposium on Applied Computing (SAC ’22), Apr 2022, Prague (virtual), Czech Republic. pp.829-836, ⟨10.1145/3477314.3507274⟩. ⟨hal-03626768⟩
69 View
333 Download

Altmetric

Share

Gmail Facebook X LinkedIn More