Interactive Discovery and Exploration of Visual Bias in Generative Text-to-Image Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Eschner, Johannes, Labadie-Tamayo, Roberto, Zeppelzauer, Matthias, Waldner, Manuela
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914396737896448
author Eschner, Johannes
Labadie-Tamayo, Roberto
Zeppelzauer, Matthias
Waldner, Manuela
author_facet Eschner, Johannes
Labadie-Tamayo, Roberto
Zeppelzauer, Matthias
Waldner, Manuela
contents Bias in generative Text-to-Image (T2I) models is a known issue, yet systematically analyzing such models' outputs to uncover it remains challenging. We introduce the Visual Bias Explorer (ViBEx) to interactively explore the output space of T2I models to support the discovery of visual bias. ViBEx introduces a novel flexible prompting tree interface in combination with zero-shot bias probing using CLIP for quick and approximate bias exploration. It additionally supports in-depth confirmatory bias analysis through visual inspection of forward, intersectional, and inverse bias queries. ViBEx is model-agnostic and publicly available. In four case study interviews, experts in AI and ethics were able to discover visual biases that have so far not been described in literature.
format Preprint
id arxiv_https___arxiv_org_abs_2504_19703
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Interactive Discovery and Exploration of Visual Bias in Generative Text-to-Image Models
Eschner, Johannes
Labadie-Tamayo, Roberto
Zeppelzauer, Matthias
Waldner, Manuela
Human-Computer Interaction
Bias in generative Text-to-Image (T2I) models is a known issue, yet systematically analyzing such models' outputs to uncover it remains challenging. We introduce the Visual Bias Explorer (ViBEx) to interactively explore the output space of T2I models to support the discovery of visual bias. ViBEx introduces a novel flexible prompting tree interface in combination with zero-shot bias probing using CLIP for quick and approximate bias exploration. It additionally supports in-depth confirmatory bias analysis through visual inspection of forward, intersectional, and inverse bias queries. ViBEx is model-agnostic and publicly available. In four case study interviews, experts in AI and ethics were able to discover visual biases that have so far not been described in literature.
title Interactive Discovery and Exploration of Visual Bias in Generative Text-to-Image Models
topic Human-Computer Interaction
url https://arxiv.org/abs/2504.19703