Towards a critical race methodology in algorithmic fairness

Hanna, Alex; Denton, Emily; Smart, Andrew; Smith-Loud, Jamila

doi:10.1145/3351095.3372826

Cited by 221 publications

(167 citation statements)

References 87 publications

(99 reference statements)

Supporting

Mentioning

161

Contrasting

Order By: Relevance

“…Our work is related in spirit to Hanna et.al, [18] as well as Benthall and Haynes [5], who both critique the usage of racial categories from the perspective of critical race theory. We draw on Scheurman et.al's [42] survey of identity classification schemes, definitions, and annotation methods in computer vision.…”

Section: Introductionmentioning

confidence: 75%

One Label, One Billion Faces

Khan

2021

Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency

View full text Add to dashboard Cite

Computer vision is widely deployed, has highly visible, societyaltering applications, and documented problems with bias and representation. Datasets are critical for benchmarking progress in fair computer vision, and often employ broad racial categories as population groups for measuring group fairness. Similarly, diversity is often measured in computer vision datasets by ascribing and counting categorical race labels. However, racial categories are ill-defined, unstable temporally and geographically, and have a problematic history of scientific use. Although the racial categories used across datasets are superficially similar, the complexity of human race perception suggests the racial system encoded by one dataset may be substantially inconsistent with another. Using the insight that a classifier can learn the racial system encoded by a dataset, we conduct an empirical study of computer vision datasets supplying categorical race labels for face images to determine the cross-dataset consistency and generalization of racial categories. We find that each dataset encodes a substantially unique racial system, despite nominally equivalent racial categories, and some racial categories are systemically less consistent than others across datasets. We find evidence that racial categories encode stereotypes, and exclude ethnic groups from categories on the basis of nonconformity to stereotypes. Representing a billion humans under one racial category may obscure disparities and create new ones by encoding stereotypes of racial systems. The difficulty of adequately converting the abstract concept of race into a tool for measuring fairness underscores the need for a method more flexible and culturally aware than racial categories. CCS CONCEPTS • Social and professional topics → Race and ethnicity; • Computing methodologies → Computer vision.

show abstract

Section: Introductionmentioning

confidence: 75%

One Label, One Billion Faces

Khan

2021

Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency

View full text Add to dashboard Cite

show abstract

“…However, other work has approached these issues more as a problem of dataset pre-processing (Calmon et al, 2017) or database repair (Salimi et al, 2020). Critics note that domain-independent approaches may fall into what Selbst et al identify as "abstraction traps" (Selbst et al, 2019, p.60), such as failing to account for the particularities of different kinds and qualities of discrimination in a given social context -a critique Hanna et al (2020) make of fairness research that treats race as a single fixed attribute. We did not ask any questions about how papers discuss de-biasing or data cleaning due to the large number of questions we were already asking and the novelty of such approaches, but these concerns are deeply related.…”

Section: Fairness Accountability and Transparency In Machine Learningmentioning

confidence: 99%

“Garbage in, garbage out” revisited: What do machine learning application papers report about human-labeled training data?

Geiger

Cope

et al. 2021

Quantitative Science Studies

View full text Add to dashboard Cite

Supervised machine learning, in which models are automatically derived from labeled training data, is only as good as the quality of that data. This study builds on prior work that investigated to what extent ‘best practices’ around labeling training data were followed in applied ML publications within a single domain (social media platforms). In this paper, we expand by studying publications that apply supervised ML in a far broader spectrum of disciplines, focusing on human-labeled data. We report to what extent a random sample of ML application papers across disciplines give specific details about whether best practices were followed, while acknowledging that a greater range of application fields necessarily produces greater diversity of labeling and annotation methods. Because much of machine learning research and education only focuses on what is done once a “ground truth” or “gold standard” of training data is available, it is especially relevant to discuss issues around the equally-important aspect of whether such data is reliable in the first place. This determination becomes increasingly complex when applied to a variety of specialized fields, as labeling can range from a task requiring little-to-no background knowledge to one that must be performed by someone with career expertise. Peer Review https://publons.com/publon/10.1162/qss_a_00144

show abstract

“…For example, Keyes [27] shows that current studies typically treat gender classification as a purely binary problem, thereby systematically leaving out and wrongly classifying transgender people. Similarly, Hanna et al [18] argue that race and ethnicity are strongly social constructs that should not be treated as objective differences between groups. This topic, typically referred to as (algorithmic) Fairness, is an active research field that aims to counter bias and discrimination in data-driven computer systems.…”

Section: Explanationmentioning

confidence: 99%

Recommenders with a Mission

Vrijenhoek

Kaya²,

Metoui

et al. 2021

Proceedings of the 2021 Conference on Human Information Interaction and Retrieval

View full text Add to dashboard Cite

News recommenders help users to find relevant online content and have the potential to fulfill a crucial role in a democratic society, directing the scarce attention of citizens towards the information that is most important to them. Simultaneously, recent concerns about so-called filter bubbles, misinformation and selective exposure are symptomatic of the disruptive potential of these digital news recommenders. Recommender systems can make or break filter bubbles, and as such can be instrumental in creating either a more closed or a more open internet. Current approaches to evaluating recommender systems are often focused on measuring an increase in user clicks and short-term engagement, rather than measuring the user's longer term interest in diverse and important information.This paper aims to bridge the gap between normative notions of diversity, rooted in democratic theory, and quantitative metrics necessary for evaluating the recommender system. We propose a set of metrics grounded in social science interpretations of diversity and suggest ways for practical implementations.

show abstract

Towards a critical race methodology in algorithmic fairness

Cited by 221 publications

References 87 publications

One Label, One Billion Faces

One Label, One Billion Faces

“Garbage in, garbage out” revisited: What do machine learning application papers report about human-labeled training data?

Recommenders with a Mission

Contact Info

Product

Resources

About