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.
%0 Conference Paper
%1 zhao2018learning
%A Zhao, Jieyu
%A Zhou, Yichao
%A Li, Zeyu
%A Wang, Wei
%A Chang, Kai-Wei
%B EMNLP
%D 2018
%K bias embedding gender inequality neutral word
%T Learning Gender-Neutral Word Embeddings
%U http://arxiv.org/abs/1809.01496
%X 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.
@inproceedings{zhao2018learning,
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.},
added-at = {2018-10-02T15:23:38.000+0200},
author = {Zhao, Jieyu and Zhou, Yichao and Li, Zeyu and Wang, Wei and Chang, Kai-Wei},
biburl = {https://www.bibsonomy.org/bibtex/235f939fa8c140fb161e40cb371904da8/schwemmlein},
booktitle = {EMNLP},
interhash = {db2991306b249cb6a8876a7c2b8325ef},
intrahash = {35f939fa8c140fb161e40cb371904da8},
keywords = {bias embedding gender inequality neutral word},
timestamp = {2018-10-02T15:24:03.000+0200},
title = {Learning Gender-Neutral Word Embeddings},
url = {http://arxiv.org/abs/1809.01496},
year = 2018
}