Geometric Origins of Bias in Deep Neural Networks: A Human Visual System Perspective
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_ | 1866911076738662400 |
|---|---|
| author | Ma, Yanbiao Liu, Bowei Zhang, Andi |
| author_facet | Ma, Yanbiao Liu, Bowei Zhang, Andi |
| contents | Bias formation in deep neural networks (DNNs) remains a critical yet poorly understood challenge, influencing both fairness and reliability in artificial intelligence systems. Inspired by the human visual system, which decouples object manifolds through hierarchical processing to achieve object recognition, we propose a geometric analysis framework linking the geometric complexity of class-specific perceptual manifolds in DNNs to model bias. Our findings reveal that differences in geometric complexity can lead to varying recognition capabilities across categories, introducing biases. To support this analysis, we present the Perceptual-Manifold-Geometry library, designed for calculating the geometric properties of perceptual manifolds. The toolkit has been downloaded and installed over 4,500 times. This work provides a novel geometric perspective on bias formation in modern learning systems and lays a theoretical foundation for developing more equitable and robust artificial intelligence. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_11809 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Geometric Origins of Bias in Deep Neural Networks: A Human Visual System Perspective Ma, Yanbiao Liu, Bowei Zhang, Andi Computer Vision and Pattern Recognition Artificial Intelligence Bias formation in deep neural networks (DNNs) remains a critical yet poorly understood challenge, influencing both fairness and reliability in artificial intelligence systems. Inspired by the human visual system, which decouples object manifolds through hierarchical processing to achieve object recognition, we propose a geometric analysis framework linking the geometric complexity of class-specific perceptual manifolds in DNNs to model bias. Our findings reveal that differences in geometric complexity can lead to varying recognition capabilities across categories, introducing biases. To support this analysis, we present the Perceptual-Manifold-Geometry library, designed for calculating the geometric properties of perceptual manifolds. The toolkit has been downloaded and installed over 4,500 times. This work provides a novel geometric perspective on bias formation in modern learning systems and lays a theoretical foundation for developing more equitable and robust artificial intelligence. |
| title | Geometric Origins of Bias in Deep Neural Networks: A Human Visual System Perspective |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2502.11809 |