Fairness-Aware Low-Rank Adaptation Under Demographic Privacy Constraints
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_ | 1866915186444599296 |
|---|---|
| author | Kamalaruban, Parameswaran Anderson, Mark Burrell, Stuart Madigan, Maeve Skalski, Piotr Sutton, David |
| author_facet | Kamalaruban, Parameswaran Anderson, Mark Burrell, Stuart Madigan, Maeve Skalski, Piotr Sutton, David |
| contents | Pre-trained foundation models can be adapted for specific tasks using Low-Rank Adaptation (LoRA). However, the fairness properties of these adapted classifiers remain underexplored. Existing fairness-aware fine-tuning methods rely on direct access to sensitive attributes or their predictors, but in practice, these sensitive attributes are often held under strict consumer privacy controls, and neither the attributes nor their predictors are available to model developers, hampering the development of fair models. To address this issue, we introduce a set of LoRA-based fine-tuning methods that can be trained in a distributed fashion, where model developers and fairness auditors collaborate without sharing sensitive attributes or predictors. In this paper, we evaluate three such methods - sensitive unlearning, adversarial training, and orthogonality loss - against a fairness-unaware baseline, using experiments on the CelebA and UTK-Face datasets with an ImageNet pre-trained ViT-Base model. We find that orthogonality loss consistently reduces bias while maintaining or improving utility, whereas adversarial training improves False Positive Rate Parity and Demographic Parity in some cases, and sensitive unlearning provides no clear benefit. In tasks where significant biases are present, distributed fairness-aware fine-tuning methods can effectively eliminate bias without compromising consumer privacy and, in most cases, improve model utility. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_05684 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Fairness-Aware Low-Rank Adaptation Under Demographic Privacy Constraints Kamalaruban, Parameswaran Anderson, Mark Burrell, Stuart Madigan, Maeve Skalski, Piotr Sutton, David Machine Learning Computer Vision and Pattern Recognition Pre-trained foundation models can be adapted for specific tasks using Low-Rank Adaptation (LoRA). However, the fairness properties of these adapted classifiers remain underexplored. Existing fairness-aware fine-tuning methods rely on direct access to sensitive attributes or their predictors, but in practice, these sensitive attributes are often held under strict consumer privacy controls, and neither the attributes nor their predictors are available to model developers, hampering the development of fair models. To address this issue, we introduce a set of LoRA-based fine-tuning methods that can be trained in a distributed fashion, where model developers and fairness auditors collaborate without sharing sensitive attributes or predictors. In this paper, we evaluate three such methods - sensitive unlearning, adversarial training, and orthogonality loss - against a fairness-unaware baseline, using experiments on the CelebA and UTK-Face datasets with an ImageNet pre-trained ViT-Base model. We find that orthogonality loss consistently reduces bias while maintaining or improving utility, whereas adversarial training improves False Positive Rate Parity and Demographic Parity in some cases, and sensitive unlearning provides no clear benefit. In tasks where significant biases are present, distributed fairness-aware fine-tuning methods can effectively eliminate bias without compromising consumer privacy and, in most cases, improve model utility. |
| title | Fairness-Aware Low-Rank Adaptation Under Demographic Privacy Constraints |
| topic | Machine Learning Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2503.05684 |