Wikipedia in the world of global gender inequality indices

Klein, Max; Konieczny, Piotr

doi:10.1145/2788993.2789849

Cited by 9 publications

(8 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We label pages with a gender based on the number of gendered pronouns (he vs. she vs. they) and label each paragraph in the page with this label for the ABOUT dimension. 4 Wikipedia is well known to have gender bias in equity of biographical coverage and lexical bias in noun references to women (Reagle and Rhue, 2011;Graells-Garrido et al, 2015;Wagner et al, 2015;Klein and Konieczny, 2015;Klein et al, 2016;Wagner et al, 2016), making it an interesting test bed for our investigation.…”

Section: A1 Existing Data Annotationmentioning

confidence: 99%

Multi-Dimensional Gender Bias Classification

Dinan¹,

Fan²,

Wu³

et al. 2020

Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)

View full text Add to dashboard Cite

Machine learning models are trained to find patterns in data. NLP models can inadvertently learn socially undesirable patterns when training on gender biased text. In this work, we propose a novel, general framework that decomposes gender bias in text along several pragmatic and semantic dimensions: bias from the gender of the person being spoken about, bias from the gender of the person being spoken to, and bias from the gender of the speaker. Using this fine-grained framework, we automatically annotate eight large scale datasets with gender information. In addition, we collect a new, crowdsourced evaluation benchmark. Distinguishing between gender bias along multiple dimensions enables us to train better and more fine-grained gender bias classifiers. We show our classifiers are valuable for a variety of applications, like controlling for gender bias in generative models, detecting gender bias in arbitrary text, and classifying text as offensive based on its genderedness.

show abstract

Section: A1 Existing Data Annotationmentioning

confidence: 99%

Multi-Dimensional Gender Bias Classification

Dinan¹,

Fan²,

Wu³

et al. 2020

Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)

View full text Add to dashboard Cite

show abstract

“…5 Other work has shown that Wikipedia's content tends to suffer a range of gender biases and gaps as well-for example, by using terms and images that tend to reflect existing gender bias. 6 Some work has also connected explanations of the gender gap among contributors to inequality and bias in articles. Existing Wikipedia communities may deter women and others from editing and may define and enforce criteria for article creation in ways that differentially impact articles about or of interest to women.…”

Section: The Gender Gapmentioning

confidence: 99%

The Most Important Laboratory for Social Scientific and Computing Research in History

Hill¹,

Shaw²

2020

Wikipedia @ 20

View full text Add to dashboard Cite

Wikipedia's founders could not have dreamed they were creating the most important laboratory for social scientific and computing research in history, yet Wikipedia has had an enormous effect on academic research. Twenty years ago, Wikipedia's founders could not have dreamed they were creating the most important laboratory for social scientific and computing research in history. And yet that is exactly what has happened. Wikipedia and its sister projects have launched a thriving scholarly literature. How thriving? Results from Google Scholar suggest that over six thousand scholarly publications mention Wikipedia in their title and over 1.7 million mention it somewhere in their text. For comparison, the phrase "Catholic church"-an organization with a nearly two-thousand-year head startreturns about the same number of mentions in publication titles. In under twenty years, Wikipedia has become one of the most heavily studied organizations of any kind. To the extent that Wikipedia research is a field of study, what major areas of investigation have been pursued in the field so far? What are the big discoveries? The most striking gaps? This essay addresses these questions and considers some of the most important directions Wikipedia research might take in the future. The State of Wikimedia Research In 2008, Mako Hill was about to start his first year as a social science graduate student at the Massachusetts Institute of Technology where he hoped to study, among other things, organizational processes that had driven Wikipedia's success. Mako felt it would behoove him to become better connected to the recent academic scholarship on Wikipedia. He was also looking for a

show abstract

“…We label pages with a gender based on the number of gendered pronouns (he vs. she vs. they) and label each paragraph in the page with this label for the ABOUT dimension. 7 Wikipedia is well known to have gender bias in equity of biographical coverage and lexical bias in noun references to women (Reagle and Rhue, 2011;Graells-Garrido et al, 2015;Wagner et al, 2015;Klein and Konieczny, 2015;Klein et al, 2016;Wagner et al, 2016), making it an interesting test bed for our investigation.…”

Section: A Existing Data Annotationmentioning

confidence: 99%

Multi-Dimensional Gender Bias Classification

Dinan¹,

Fan²,

Wu³

et al. 2020

Preprint

View full text Add to dashboard Cite

Machine learning models are trained to find patterns in data. NLP models can inadvertently learn socially undesirable patterns when training on gender biased text. In this work, we propose a general framework that decomposes gender bias in text along several pragmatic and semantic dimensions: bias from the gender of the person being spoken about, bias from the gender of the person being spoken to, and bias from the gender of the speaker. Using this fine-grained framework, we automatically annotate eight large scale datasets with gender information. In addition, we collect a novel, crowdsourced evaluation benchmark of utterance-level gender rewrites. Distinguishing between gender bias along multiple dimensions is important, as it enables us to train finer-grained gender bias classifiers. We show our classifiers prove valuable for a variety of important applications, such as controlling for gender bias in generative models, detecting gender bias in arbitrary text, and shed light on offensive language in terms of genderedness.

show abstract

Wikipedia in the world of global gender inequality indices

Cited by 9 publications

References 2 publications

Multi-Dimensional Gender Bias Classification

Multi-Dimensional Gender Bias Classification

The Most Important Laboratory for Social Scientific and Computing Research in History

Multi-Dimensional Gender Bias Classification

Contact Info

Product

Resources

About