Survey of Bias In Text-to-Image Generation: Definition, Evaluation, and Mitigation
read the original abstract
The recent advancement of large and powerful models with Text-to-Image (T2I) generation abilities -- such as OpenAI's DALLE-3 and Google's Gemini -- enables users to generate high-quality images from textual prompts. However, it has become increasingly evident that even simple prompts could cause T2I models to exhibit conspicuous social bias in generated images. Such bias might lead to both allocational and representational harms in society, further marginalizing minority groups. Noting this problem, a large body of recent works has been dedicated to investigating different dimensions of bias in T2I systems. However, an extensive review of these studies is lacking, hindering a systematic understanding of current progress and research gaps. We present the first extensive survey on bias in T2I generative models. In this survey, we review prior studies on dimensions of bias: Gender, Skintone, and Geo-Culture. Specifically, we discuss how these works define, evaluate, and mitigate different aspects of bias. We found that: (1) while gender and skintone biases are widely studied, geo-cultural bias remains under-explored; (2) most works on gender and skintone bias investigated occupational association, while other aspects are less frequently studied; (3) almost all gender bias works overlook non-binary identities in their studies; (4) evaluation datasets and metrics are scattered, with no unified framework for measuring biases; and (5) current mitigation methods fail to resolve biases comprehensively. Based on current limitations, we point out future research directions that contribute to human-centric definitions, evaluations, and mitigation of biases. We hope to highlight the importance of studying biases in T2I systems, as well as encourage future efforts to holistically understand and tackle biases, building fair and trustworthy T2I technologies for everyone.
This paper has not been read by Pith yet.
Forward citations
Cited by 9 Pith papers
-
Gender Artifacts from Art History to Text-to-Image Generation
Introduces the StyleGender dataset and PixelSGA/MaskSGA metrics showing that text-to-image models amplify gender artifacts present in artistic styles beyond historical baselines.
-
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...
-
Embedding Arithmetic: A Lightweight, Tuning-Free Framework for Post-hoc Bias Mitigation in Text-to-Image Models
Embedding Arithmetic performs vector operations in the embedding space of T2I models to mitigate bias at inference time, outperforming baselines on diversity while preserving coherence via a new Concept Coherence Score.
-
BiasIG: Benchmarking Multi-dimensional Social Biases in Text-to-Image Models
BiasIG is a multi-dimensional benchmark for social biases in T2I models that shows debiasing interventions frequently cause confounding discrimination effects.
-
Controllable Image Generation with Composed Parallel Token Prediction
A new formulation for composing discrete generative processes enables precise control over novel condition combinations in image generation, cutting error rates by 63% and speeding up inference.
-
VideoPhy: Evaluating Physical Commonsense for Video Generation
VideoPhy benchmark shows state-of-the-art text-to-video models follow physical commonsense and text prompts in only 39.6% of cases for the best model.
-
Controllable Image Generation with Composed Parallel Token Prediction
A derived formulation for composing discrete probabilistic generative processes enables novel condition combinations in image generation, yielding 63.4% relative error reduction and FID gains on CLEVR and FFHQ datasets.
-
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.
-
Beyond Categories of Caste: Examining Caste Bias and Morality in Text-to-Image AI Models
The paper reframes caste as relational rather than categorical and combines algorithmic audit with critical discourse analysis to examine nuanced caste biases in T2I models while proposing an anti-caste framework for ...
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.