Beyond Performance Disparities: A Three-Level Audit of Representational Harm in CelebA

TL;DR

This paper conducts a comprehensive three-level audit of CelebA, revealing how cultural biases related to gender and age are embedded in dataset labels, learned features, and model attention, leading to representational harms.

Contribution

It introduces a novel three-level analysis of biases in CelebA, highlighting how cultural stereotypes influence dataset structure, feature weights, and model focus, and discusses implications for fairness research.

Findings

Hierarchical clustering reveals gendered archetypes aligned with cultural stereotypes.

Model attention varies by gender and age, reflecting societal double standards.

Older males have high accuracy but low precision, indicating categorical exclusion.

Abstract

Large-scale facial datasets like CelebA are widely used in computer vision, yet the cultural biases embedded in their labels remain underexplored. Fairness research has distinguished representational from allocational harms, but audits of computer vision datasets have mostly examined categorical labels, leaving open how such harms appear in learned features and model attention. This paper examines CelebA at three levels: dataset structure, learned feature weights, and spatial attention, focusing on how gendered double standards of ageing and beauty are encoded in the data and reproduced in model behaviour. First, hierarchical clustering of 202,599 images shows that the 39 attributes organise into latent trait bundles aligned with cultural archetypes: performative femininity (youth, makeup, adornment) and professional masculinity (ageing, facial hair, formal attire). Female faces, though more often rated attractive overall, incur steep penalties when assigned to ageing or masculine-coded clusters. Second, XGBoost with SHAP analysis reveal gender-specific effects, such as adiposity reducing attractiveness only for females. Third, Grad-CAM finds that predictions for female and younger male subgroups concentrate on mid-face cues, whereas predictions for older males drift toward peripheral cues such as hair and clothing. Older males attain the highest accuracy but the lowest average precision, indicating categorical exclusion of groups outside the dataset's evaluative templates. Cultural double standards thus pass from media representation into dataset labels, feature weights, and model attention, producing two representational harms: hyper-scrutiny of women under a narrow evaluative template, and exclusion of older men from the scheme entirely. Fairness metrics focused on performance disparities mask both, underscoring the need to address representational harm in fairness research.