Equalizing Gender Biases in Neural Machine Translation with Word Embeddings Techniques

TL;DR

This paper proposes a method to reduce gender bias in neural machine translation by applying debiasing techniques to word embeddings within the Transformer model, leading to fairer translations and improved BLEU scores.

Contribution

It introduces a novel approach to mitigate gender bias in neural machine translation by integrating debiasing techniques on word embeddings in the Transformer architecture.

Findings

Achieved up to one BLEU point improvement.

Successfully learned to equalize existing gender biases.

Demonstrated effectiveness on WMT English-Spanish benchmark.

Abstract

Neural machine translation has significantly pushed forward the quality of the field. However, there are remaining big issues with the output translations and one of them is fairness. Neural models are trained on large text corpora which contain biases and stereotypes. As a consequence, models inherit these social biases. Recent methods have shown results in reducing gender bias in other natural language processing tools such as word embeddings. We take advantage of the fact that word embeddings are used in neural machine translation to propose a method to equalize gender biases in neural machine translation using these representations. Specifically, we propose, experiment and analyze the integration of two debiasing techniques over GloVe embeddings in the Transformer translation architecture. We evaluate our proposed system on the WMT English-Spanish benchmark task, showing gains up to one BLEU point. As for the gender bias evaluation, we generate a test set of occupations and we show that our proposed system learns to equalize existing biases from the baseline system.