paper-with-me

홈 › Papers

Hidden Bias in the Machine: Stereotypes in Text-to-Image Models

2025-06-09 · Sedat Porikli, Vedat Porikli

Text-to-Image (T2I) models have transformed visual content creation, producing highly realistic images from natural language prompts. However, concerns persist around their potential to replicate and magnify existing societal biases. To investigate these issues, we curated a diverse set of prompts spanning thematic categories such as occupations, traits, actions, ideologies, emotions, family roles, place descriptions, spirituality, and life events. For each of the 160 unique topics, we crafted multiple prompt variations to reflect a wide range of meanings and perspectives. Using Stable Diffusion 1.5 (UNet-based) and Flux-1 (DiT-based) models with original checkpoints, we generated over 16,000 images under consistent settings. Additionally, we collected 8,000 comparison images from Google Image Search. All outputs were filtered to exclude abstract, distorted, or nonsensical results. Our analysis reveals significant disparities in the representation of gender, race, age, somatotype, and other human-centric factors across generated images. These disparities often mirror and reinforce harmful stereotypes embedded in societal narratives. We discuss the implications of these findings and emphasize the need for more inclusive datasets and development practices to foster fairness in generative visual systems.

📄 PDF Abstract BibTeX arXiv:2506.13780

Code (0)

등록된 구현이 없습니다.

Tasks

FairnessImage Retrieval

Methods 이 논문이 사용한 방법론

Diffusion Diffusion models generate samples by gradually removing noise from a signal, and their training objective can be expressed as a reweighted variational lower-bound…
SET Dynamic Sparse Training method where weight mask is updated randomly periodically

Similar Papers 제목 키워드 기반

The Hidden Bias: A Study on Explicit and Implicit Political Stereotypes in Large Language Models

2025-10-09 · Konrad Löhr, Shuzhou Yuan, Michael Färber arxiv

Large Language Models (LLMs) are increasingly integral to information dissemination and decision-making processes. Given their growing societal influence, understanding potential biases, particularly within the political…

Easily Accessible Text-to-Image Generation Amplifies Demographic Stereotypes at Large Scale

2022-11-07 · Federico Bianchi, Pratyusha Kalluri, Esin Durmus, Faisal Ladhak 외

Machine learning models that convert user-written text descriptions into images are now widely available online and used by millions of users to generate millions of images a day. We investigate the potential for these m…

Image GenerationText to Image GenerationText-to-Image Generation

OASIS Uncovers: High-Quality T2I Models, Same Old Stereotypes

2025-01-01 · Sepehr Dehdashtian, Gautam Sreekumar, Vishnu Naresh Boddeti

Images generated by text-to-image (T2I) models often exhibit visual biases and stereotypes of concepts such as culture and profession. Existing quantitative measures of stereotypes are based on statistical parity that do…

AttributeImage Generation

"Im not Racist but...": Discovering Bias in the Internal Knowledge of Large Language Models

2023-10-13 · Abel Salinas, Louis Penafiel, Robert McCormack, Fred Morstatter

Large language models (LLMs) have garnered significant attention for their remarkable performance in a continuously expanding set of natural language processing tasks. However, these models have been shown to harbor inhe…

Fairness

Quantifying and Reducing Stereotypes in Word Embeddings

2016-06-20 · Tolga Bolukbasi, Kai-Wei Chang, James Zou, Venkatesh Saligrama 외

Machine learning algorithms are optimized to model statistical properties of the training data. If the input data reflects stereotypes and biases of the broader society, then the output of the learning algorithm also cap…

Word Embeddings