Equalizing Gender Bias in Neural Machine Translation with Word Embeddings Techniques

By Joel Escudé Font, Marta Ruiz Costa-jussà. 2019

182 citations recorded by the source corpus.

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.

Subfields: Artificial Intelligence, Natural Language Processing, Safety, Robustness & Interpretability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Natural Language Processing

In Safety, Robustness & Interpretability

Citations in the corpus

Language (Technology) is Power: A Critical Survey of “Bias” in NLP (2020, background): “…(2019); Cao and Daumé (2019); Cho et al. (2019); Davidson et al. (2019); Dev et al. (2019); Dinan et al. (2019); Fisher (2019); Florez (2019); Font and Costa-jussà (2019); Garg et al. (2019); Huang et al. (2019); Liu et al. (2019); Nozza et al. (2019); Prabhakaran et al. (2019); Qian et al.…”
Toward Gender-Inclusive Coreference Resolution (2019, background): “, 2019; Gonen and Goldberg, 2019), sentiment analysis (Kiritchenko and Mohammad, 2018), machine translation (Font and Costa-jussà, 2019; Prates et al., 2019; Dryer, 2013; Frank et al., 2004; Wandruszka, 1969; Nissen, 2002; Doleschal and Schmid, 2001), among many”
Evaluating Gender Bias in Machine Translation (2019, background): “, translating nurses as females and programmers as males, regardless of context (Alvarez-Melis and Jaakkola, 2017; Font and Costa-Jussà, 2019).”
Gender Bias in Contextualized Word Embeddings (2019, background): “, 2018), machine translation (Font and Costa-jussà, 2019; Vanmassenhove et al., 2018), and web search (Kay et al.”
Evaluating the Underlying Gender Bias in Contextualized Word Embeddings (2019, background): “Examples in the area of NLP are the case of machine translation that systems tend to ignore the coreference information in benefit of an stereotype (Font and Costa-jussà, 2019) or sentiment analysis where higher sentiment intensity prediction is biased for a particular gender (Kiritchenko and Mohamm”
A Survey on Bias and Fairness in Machine Learning (2019, methodology): “To solve this, authors in [45] take advantage of the fact that word embeddings are used in machine translation.”
Lawyers are Dishonest? Quantifying Representational Harms in Commonsense Knowledge Resources (2021, background): “, 2020), machine translation (Font and Costa-jussà, 2019), word embedding (Bolukbasi et al.”
Man is to Person as Woman is to Location: Measuring Gender Bias in Named Entity Recognition (2019, background): “2018a), machine translation (Font and Costa-jussà 2019) and sentence encoders (May et al.”
You reap what you sow: On the Challenges of Bias Evaluation Under Multilingual Settings (2022, background): “For instance, many gender bias evaluations focus on Western professions, which are grammatically gendered in some languages (Chen et al., 2021; Zhou et al., 2019) or may not cover other prevalent occupations outside the U.S. (Escudé Font and Costa-jussà, 2019).”
Examining Gender Bias in Languages with Grammatical Gender (2019, background): “Bias in word embedding may affect the downstream applications (Zhao et al., 2018a; Rudinger et al., 2018; Font and Costa-jussà, 2019).”
Automatic Gender Identification and Reinflection in Arabic (2019, methodology): “More recently, Font and Costa-jussà (2019) studied the impact of gender debiasing on NMT between English and Spanish using debi-ased and gender-neutral word embeddings.”
Towards Understanding Gender Bias in Relation Extraction (2019, background): “, 2018) and this can affect bias in downstream predictions for models using the embeddings (Zhao et al., 2018; Font and Costa-Jussa, 2019).”
exBERT: A Visual Analysis Tool to Explore Learned Representations in Transformer Models (2020, methodology): “, 2018), a problem inherent in the training data that infects the model’s understanding of language (Font and Costa-jussà, 2019).”