Person-Centric Annotations of LAION-400M: Auditing Bias and Its Transfer to Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908918354018304 |
|---|---|
| author | Girrbach, Leander Alaniz, Stephan Smith, Genevieve Darrell, Trevor Akata, Zeynep |
| author_facet | Girrbach, Leander Alaniz, Stephan Smith, Genevieve Darrell, Trevor Akata, Zeynep |
| contents | Vision-language models trained on large-scale multimodal datasets show strong demographic biases, but the role of training data in producing these biases remains unclear. A major barrier has been the lack of demographic annotations in web-scale datasets such as LAION-400M. We address this gap by creating person-centric annotations for the full dataset, including over 276 million bounding boxes, perceived gender and race/ethnicity labels, and automatically generated captions. These annotations are produced through validated automatic labeling pipelines combining object detection, multimodal captioning, and finetuned classifiers. Using them, we uncover demographic imbalances and harmful associations, such as the disproportionate linking of men and individuals perceived as Black or Middle Eastern with crime-related and negative content. We also show that a linear fit predicts 60-70% of gender bias in CLIP and Stable Diffusion from direct co-occurrences in the data. Our resources establish the first large-scale empirical link between dataset composition and downstream model bias. Code is available at https://github.com/ExplainableML/LAION-400M-Person-Centric-Annotations. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_03721 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Person-Centric Annotations of LAION-400M: Auditing Bias and Its Transfer to Models Girrbach, Leander Alaniz, Stephan Smith, Genevieve Darrell, Trevor Akata, Zeynep Computer Vision and Pattern Recognition Computation and Language Computers and Society Machine Learning Vision-language models trained on large-scale multimodal datasets show strong demographic biases, but the role of training data in producing these biases remains unclear. A major barrier has been the lack of demographic annotations in web-scale datasets such as LAION-400M. We address this gap by creating person-centric annotations for the full dataset, including over 276 million bounding boxes, perceived gender and race/ethnicity labels, and automatically generated captions. These annotations are produced through validated automatic labeling pipelines combining object detection, multimodal captioning, and finetuned classifiers. Using them, we uncover demographic imbalances and harmful associations, such as the disproportionate linking of men and individuals perceived as Black or Middle Eastern with crime-related and negative content. We also show that a linear fit predicts 60-70% of gender bias in CLIP and Stable Diffusion from direct co-occurrences in the data. Our resources establish the first large-scale empirical link between dataset composition and downstream model bias. Code is available at https://github.com/ExplainableML/LAION-400M-Person-Centric-Annotations. |
| title | Person-Centric Annotations of LAION-400M: Auditing Bias and Its Transfer to Models |
| topic | Computer Vision and Pattern Recognition Computation and Language Computers and Society Machine Learning |
| url | https://arxiv.org/abs/2510.03721 |