Unequal Representation and Gender Stereotypes in Image Search Results for Occupations

By Matthew Kay, Cynthia Matuszek, Sean A Munson. 2015

573 citations recorded by the source corpus.

Subfields: Gender Studies, Sociology and Political Science, 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 Gender Studies

In Sociology and Political Science

In Safety, Robustness & Interpretability

Citations in the corpus

Toward Gender-Inclusive Coreference Resolution (2019, background): “we focus largely on harms due to over- and underrepresentation (Kay et al., 2015), replicating stereotypes (Sweeney, 2013; Caliskan et al.”
Diagnosing Gender Bias in Image Recognition Systems (2020, background): “For example, image search algorithms, when asked to return images for occupations, generated results that reproduced gendered stereotypes, exaggerating gender disparities (Kay, Matuszek, and Munson 2015) and featuring women less prominently than men (Lam et al. 2018).”
Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings (2016, background): “In Wikipedia, Wager et al. [39] found that, as suggested by prior work on gender bias in language [2], articles about women more often emphasize their gender, their husbands and their husbands’ jobs, and other topics discussed consistently less often than in articles about men.”
Understanding and Evaluating Racial Biases in Image Captioning (2021, background): “For example, gender and racial biases are prevalent in image tagging [62, 7] and image search [39, 52]; visual recognition models have disparate error rates across demographics and geographic regions [16, 24].”
Does Object Recognition Work for Everyone? (2019, neutral): “A study of Google Image Search revealed exaggeration of stereotypes and systematic underrepresentation of women in search results [26], and a study of ImageNet revealed correlations between classes and race [40].”
Towards fairer datasets: filtering and balancing the distribution of the people subtree in the ImageNet hierarchy (2019, methodology): “ImageNet consists of Internet images collected by querying image search engines [18], which have been demonstrated to retrieve biased results in terms of race and gender [14, 41, 55].”
Men Also Like Shopping: Reducing Gender Bias Amplification using Corpus-level Constraints (2017, background): “Such issues have been demonstrated in various learning systems, including online advertisement systems (Sweeney, 2013), word embedding models (Bolukbasi et al., 2016; Caliskan et al., 2017), online news (Ross and Carter, 2011), web search (Kay et al., 2015), and credit score (Hardt et al., 2016).”
Gender Bias in Coreference Resolution: Evaluation and Debiasing Methods (2018, background): “, 2017), online news (Ross and Carter, 2011), web search (Kay et al., 2015) and advertisements (Sweeney, 2013).”
Learning Gender-Neutral Word Embeddings (2018, background): “Stereotype Analysis Implicit stereotypes have been observed in applications such as on-line advertising systems (Sweeney, 2013), web search (Kay et al., 2015), and online reviews (Wal-lace and Paul, 2016).”
Gender Bias in Contextualized Word Embeddings (2019, neutral): “…applications relying on automatic language analysis, including online news (Ross and Carter, 2011), advertisements (Sweeney, 2013), abusive language detection (Park et al., 2018), machine translation (Font and Costa-juss`a, 2019; Vanmassen-hove et al., 2018), and web search (Kay et al., 2015).”
On formalizing fairness in prediction with machine learning (2017, background): “ of the modern computing devices to swi›ly replicate decision-making can lead to magnifying erroneous behavior [9]. Barocas and Selbst [6], Citron and Pasquale [12], Edelman and Luca [24], Kay et al. [42] and Sweeney [65] show that algorithms can make discriminatory decisions even if the computing p”
The Price of Fair PCA: One Extra Dimension (2018, background): “…labeled African Americans as gorillas [Twitter, 2015; Simonite, 2018] and returned queries for CEOs with images overwhelmingly male and white [Kay et al., 2015], searches for African American names caused the display of arrest record advertisements with higher frequency than searches for…”
Improving Fairness in Machine Learning Systems: What Do Industry Practitioners Need? (2018, background): “Other work has focused on auditing widely-used ML products from the outside [8, 19, 26, 58], often coupled with high-level calls to action aimed at those responsible for developing and maintaining these systems or for regulating their use [33, 86, 88].”
Large image datasets: A pyrrhic win for computer vision? (2020, positive): “For example, image search results both exaggerate stereotypes and systematically under-represent women in search results for occupations [47]; object detection systems designed to detect pedestrians display higher error rates for recognition of demographic groups with dark skin tones [87]; and…”
Examining Gender Bias in Languages with Grammatical Gender (2019, background): “applications, such as online reviews (Wallace and Paul, 2016), advertisement (Sweeney, 2013) and web search (Kay et al., 2015).”
Ranking with Fairness Constraints (2017, background): “From search results that inadvertently promote stereotypes by over/under-representing sensitive attributes such as race and gender [KMM15, BCZ + 16], to news feeds that can promote extremist ideology [CHRG16] and possibly even aﬀect the results of elections [Bae16, BMA15], it is clear that ensuring…”
Investigating Societal Biases in a Poetry Composition System (2020, background): “Earlier work on biases in information retrieval systems include examining biases in resume search engines (Chen et al., 2018) and image search results (Kay et al., 2015).”
Algorithmic Fairness (2020, background): “[82] show that image searches of occupations in Google’s engine resulted in gender-biased results.”