CIDER: A Causal Cure for Brand-Obsessed Text-to-Image Models

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Shen, Fangjian, Liang, Zifeng, Wang, Chao, Wen, Wushao
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866908547160211456
author Shen, Fangjian
Liang, Zifeng
Wang, Chao
Wen, Wushao
author_facet Shen, Fangjian
Liang, Zifeng
Wang, Chao
Wen, Wushao
contents Text-to-image (T2I) models exhibit a significant yet under-explored "brand bias", a tendency to generate contents featuring dominant commercial brands from generic prompts, posing ethical and legal risks. We propose CIDER, a novel, model-agnostic framework to mitigate bias at inference-time through prompt refinement to avoid costly retraining. CIDER uses a lightweight detector to identify branded content and a Vision-Language Model (VLM) to generate stylistically divergent alternatives. We introduce the Brand Neutrality Score (BNS) to quantify this issue and perform extensive experiments on leading T2I models. Results show CIDER significantly reduces both explicit and implicit biases while maintaining image quality and aesthetic appeal. Our work offers a practical solution for more original and equitable content, contributing to the development of trustworthy generative AI.
format Preprint
id arxiv_https___arxiv_org_abs_2509_15803
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CIDER: A Causal Cure for Brand-Obsessed Text-to-Image Models
Shen, Fangjian
Liang, Zifeng
Wang, Chao
Wen, Wushao
Computer Vision and Pattern Recognition
Artificial Intelligence
Text-to-image (T2I) models exhibit a significant yet under-explored "brand bias", a tendency to generate contents featuring dominant commercial brands from generic prompts, posing ethical and legal risks. We propose CIDER, a novel, model-agnostic framework to mitigate bias at inference-time through prompt refinement to avoid costly retraining. CIDER uses a lightweight detector to identify branded content and a Vision-Language Model (VLM) to generate stylistically divergent alternatives. We introduce the Brand Neutrality Score (BNS) to quantify this issue and perform extensive experiments on leading T2I models. Results show CIDER significantly reduces both explicit and implicit biases while maintaining image quality and aesthetic appeal. Our work offers a practical solution for more original and equitable content, contributing to the development of trustworthy generative AI.
title CIDER: A Causal Cure for Brand-Obsessed Text-to-Image Models
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2509.15803