REVIEW 11 cited by
Stable Bias: Analyzing Societal Representations in Diffusion Models
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Stable Bias: Analyzing Societal Representations in Diffusion Models
read the original abstract
As machine learning-enabled Text-to-Image (TTI) systems are becoming increasingly prevalent and seeing growing adoption as commercial services, characterizing the social biases they exhibit is a necessary first step to lowering their risk of discriminatory outcomes. This evaluation, however, is made more difficult by the synthetic nature of these systems' outputs: common definitions of diversity are grounded in social categories of people living in the world, whereas the artificial depictions of fictive humans created by these systems have no inherent gender or ethnicity. To address this need, we propose a new method for exploring the social biases in TTI systems. Our approach relies on characterizing the variation in generated images triggered by enumerating gender and ethnicity markers in the prompts, and comparing it to the variation engendered by spanning different professions. This allows us to (1) identify specific bias trends, (2) provide targeted scores to directly compare models in terms of diversity and representation, and (3) jointly model interdependent social variables to support a multidimensional analysis. We leverage this method to analyze images generated by 3 popular TTI systems (Dall-E 2, Stable Diffusion v 1.4 and 2) and find that while all of their outputs show correlations with US labor demographics, they also consistently under-represent marginalized identities to different extents. We also release the datasets and low-code interactive bias exploration platforms developed for this work, as well as the necessary tools to similarly evaluate additional TTI systems.
Forward citations
Cited by 11 Pith papers
-
Do Counterfactually Fair Image Classifiers Satisfy Group Fairness? -- A Theoretical and Empirical Study
Counterfactually fair image classifiers need not satisfy group fairness when a latent attribute G correlates with the sensitive attribute; reducing reliance on G (via CKD) restores both.
-
Efficient bias mitigation in T2I diffusion models using Concept Graphs
Joint concept-graph alignment of Stable Diffusion’s text encoder and denoiser cuts fairness discrepancy ~30%, incoherent outputs 88%, and improves adversarial unlearning robustness.
-
Toward Calibrated, Fair, and accurate Deepfake Detection
Face-Feature Tuning is a label-free logit remapping method that reduces FPR/TPR gaps across groups in deepfake detection while preserving overall accuracy.
-
Collective Recourse for Generative Urban Visualizations
Collective recourse formalizes community reports to fix group harms in diffusion models for urban visualizations via a report-triage-fix-verify pipeline, four primitives, a mandate score, and synthetic evaluation of 2...
-
EquiSteer: Cross-Attention Steering Towards a Fairer Text-Guided Image Generation
EquiSteer reduces average demographic parity gaps by up to 87% in Stable Diffusion variants and SANA via inference-time cross-attention steering with minimal impact on image quality or alignment.
-
Unmasking LAION-5B: Age, Gender, Race, and Emotion Biases in Large-Scale Image Datasets
Empirical audit of LAION-2B-en and LAION-2B-multi finds overrepresentation of young adults, White people, and males plus stereotypical emotion associations across two attribute classifiers.
-
BAFIS: Dataset + Framework to assess occupational Bias and Human Preference in modern Text-to-image Models
BAFIS supplies a new dataset and human-feedback framework demonstrating systematic gender and ethnicity biases in occupational image generation by five text-to-image models, with partial alignment between automated me...
-
Mechanisms of Misgeneralization in Physical Sequence Modeling
Generative sequence models for physical tasks exhibit physical misgeneralization where local prediction errors propagate through physical measurements to distort aggregate distributions over quantities like distance o...
-
Closed-Form Concept Erasure via Double Projections
A training-free double-projection linear transformation erases target concepts from generative models by computing a proxy projection then applying a constrained update in the left null space of known directions.
-
Who Defines Fairness? Target-Based Prompting for Demographic Representation in Generative Models
Target-based prompting lets users define fairness distributions for skin tones in generative AI, shifting outputs closer to chosen targets across 36 tested prompts for occupations and contexts.
-
T2I-BiasBench: A Multi-Metric Framework for Auditing Demographic and Cultural Bias in Text-to-Image Models
T2I-BiasBench is a new 13-metric benchmark showing that text-to-image models amplify gender and beauty biases while collapsing cultural representations even in aligned systems like Gemini.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.