English

Learning Gender-Neutral Word Embeddings

Computation and Language 2018-09-06 v1 Machine Learning Machine Learning

Abstract

Word embedding models have become a fundamental component in a wide range of Natural Language Processing (NLP) applications. However, embeddings trained on human-generated corpora have been demonstrated to inherit strong gender stereotypes that reflect social constructs. To address this concern, in this paper, we propose a novel training procedure for learning gender-neutral word embeddings. Our approach aims to preserve gender information in certain dimensions of word vectors while compelling other dimensions to be free of gender influence. Based on the proposed method, we generate a Gender-Neutral variant of GloVe (GN-GloVe). Quantitative and qualitative experiments demonstrate that GN-GloVe successfully isolates gender information without sacrificing the functionality of the embedding model.

Keywords

Cite

@article{arxiv.1809.01496,
  title  = {Learning Gender-Neutral Word Embeddings},
  author = {Jieyu Zhao and Yichao Zhou and Zeyu Li and Wei Wang and Kai-Wei Chang},
  journal= {arXiv preprint arXiv:1809.01496},
  year   = {2018}
}

Comments

EMNLP 2018

R2 v1 2026-06-23T03:55:04.801Z