Biased Embeddings from Wild Data: Measuring, Understanding and Removing

Sutton, Andrew M.; Lansdall-Welfare, Thomas; Cristianini, Nello

doi:10.1007/978-3-030-01768-2_27

Cited by 15 publications

(13 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This imbalance in representation can have a dramatic downstream effect on NLP systems trained on such a corpus, such as giving preference to male engineers over female engineers in an automated resumé filtering system. Gender stereotypes of this sort have been observed in word embeddings (Bolukbasi et al, 2016;Sutton et al, 2018), contextual word embeddings (Zhao et al, 2019), and co-reference resolution systems (Rudinger et al, 2018;Zhao et al, 2018) inter alia.…”

Section: Gender Stereotypes In Textmentioning

confidence: 99%

“…This is because NLP systems depend on language corpora, which are inherently "not objective; they are creations of human design" (Crawford, 2013). One type of societal bias that has received considerable attention from the NLP community is gender stereotyping (Garg et al, 2017;Rudinger et al, 2017;Sutton et al, 2018). Gender stereotypes can manifest in language in overt ways.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Counterfactual Data Augmentation for Mitigating Gender Stereotypes in Languages with Rich Morphology

Zmigrod¹,

Mielke²,

Wallach³

et al. 2019

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics

158

122

View full text Add to dashboard Cite

Gender stereotypes are manifest in most of the world's languages and are consequently propagated or amplified by NLP systems. Although research has focused on mitigating gender stereotypes in English, the approaches that are commonly employed produce ungrammatical sentences in morphologically rich languages. We present a novel approach for converting between masculine-inflected and feminineinflected sentences in such languages. For Spanish and Hebrew, our approach achieves F 1 scores of 82% and 73% at the level of tags and accuracies of 90% and 87% at the level of forms. By evaluating our approach using four different languages, we show that, on average, it reduces gender stereotyping by a factor of 2.5 without any sacrifice to grammaticality.

show abstract

Section: Gender Stereotypes In Textmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Counterfactual Data Augmentation for Mitigating Gender Stereotypes in Languages with Rich Morphology

Zmigrod¹,

Mielke²,

Wallach³

et al. 2019

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics

158

122

View full text Add to dashboard Cite

show abstract

“…Gender bias has been detected, studied, and partially addressed for standard and contextualized word embeddings in a number of studies (Bolukbasi et al, 2016;Caliskan et al, 2017;Sutton et al, 2018;Basta et al, 2019;Garg et al, 2018;Zhao et al, 2018Zhao et al, , 2019. These studies showed that training word embeddings on large human produced corpora such as news text leads to encoding societal biases including gender and race.…”

Section: Related Workmentioning

confidence: 99%

Automatic Gender Identification and Reinflection in Arabic

Habash¹,

Bouamor²,

Chung³

2019

Proceedings of the First Workshop on Gender Bias in Natural Language Processing

View full text Add to dashboard Cite

The impressive progress in many Natural Language Processing (NLP) applications has increased the awareness of some of the biases these NLP systems have with regards to gender identities. In this paper, we propose an approach to extend biased single-output genderblind NLP systems with gender-specific alternative reinflections. We focus on Arabic, a gender-marking morphologically rich language, in the context of machine translation (MT) from English, and for first-personsingular constructions only. Our contributions are the development of a system-independent gender-awareness wrapper, and the building of a corpus for training and evaluating firstperson-singular gender identification and reinflection in Arabic. Our results successfully demonstrate the viability of this approach with 8% relative increase in BLEU score for firstperson-singular feminine, and 5.3% comparable increase for first-person-singular masculine on top of a state-of-the-art gender-blind MT system on a held-out test set.

show abstract

“…These algorithms extract not only semantic and syntactic information from everyday language, but also subtle biases in the usage of words [41]. For example, the words nurse or housekeeper frequently take a more feminine position in the semantic space than the words pilot or engineer, which take a more masculine position [41,42].…”

Section: Introductionmentioning

confidence: 99%

English colour terms carry gender and valence biases: A corpus study using word embeddings

et al. 2021

Self Cite

View full text Add to dashboard Cite

In Western societies, the stereotype prevails that pink is for girls and blue is for boys. A third possible gendered colour is red. While liked by women, it represents power, stereotypically a masculine characteristic. Empirical studies confirmed such gendered connotations when testing colour-emotion associations or colour preferences in males and females. Furthermore, empirical studies demonstrated that pink is a positive colour, blue is mainly a positive colour, and red is both a positive and a negative colour. Here, we assessed if the same valence and gender connotations appear in widely available written texts (Wikipedia and newswire articles). Using a word embedding method (GloVe), we extracted gender and valence biases for blue, pink, and red, as well as for the remaining basic colour terms from a large English-language corpus containing six billion words. We found and confirmed that pink was biased towards femininity and positivity, and blue was biased towards positivity. We found no strong gender bias for blue, and no strong gender or valence biases for red. For the remaining colour terms, we only found that green, white, and brown were positively biased. Our finding on pink shows that writers of widely available English texts use this colour term to convey femininity. This gendered communication reinforces the notion that results from research studies find their analogue in real word phenomena. Other findings were either consistent or inconsistent with results from research studies. We argue that widely available written texts have biases on their own, because they have been filtered according to context, time, and what is appropriate to be reported.

show abstract

Biased Embeddings from Wild Data: Measuring, Understanding and Removing

Cited by 15 publications

References 19 publications

Counterfactual Data Augmentation for Mitigating Gender Stereotypes in Languages with Rich Morphology

Counterfactual Data Augmentation for Mitigating Gender Stereotypes in Languages with Rich Morphology

Automatic Gender Identification and Reinflection in Arabic

English colour terms carry gender and valence biases: A corpus study using word embeddings

Contact Info

Product

Resources

About