HiLoRA: Hierarchical Low-Rank Adaptation for Personalized Federated Learning
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912939637735424 |
|---|---|
| author | Peng, Zihao Zou, Nan Zeng, Jiandian Li, Guo Chen, Ke Li, Boyuan Wang, Tian |
| author_facet | Peng, Zihao Zou, Nan Zeng, Jiandian Li, Guo Chen, Ke Li, Boyuan Wang, Tian |
| contents | Vision Transformers (ViTs) have been widely adopted in vision tasks due to their strong transferability. In Federated Learning (FL), where full fine-tuning is communication heavy, Low-Rank Adaptation (LoRA) provides an efficient and communication-friendly way to adapt ViTs. However, existing LoRA-based federated tuning methods overlook latent client structures in real-world settings, limiting shared representation learning and hindering effective adaptation to unseen clients. To address this, we propose HiLoRA, a hierarchical LoRA framework that places adapters at three levels: root, cluster, and leaf, each designed to capture global, subgroup, and client-specific knowledge, respectively. Through cross-tier orthogonality and cascaded optimization, HiLoRA separates update subspaces and aligns each tier with its residual personalized objective. In particular, we develop a LoRA-Subspace Adaptive Clustering mechanism that infers latent client groups via subspace similarity analysis, thereby facilitating knowledge sharing across structurally aligned clients. Theoretically, we establish a tier-wise generalization analysis that supports HiLoRA's design. Experiments on ViT backbones with CIFAR-100 and DomainNet demonstrate consistent improvements in both personalization and generalization. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_02785 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | HiLoRA: Hierarchical Low-Rank Adaptation for Personalized Federated Learning Peng, Zihao Zou, Nan Zeng, Jiandian Li, Guo Chen, Ke Li, Boyuan Wang, Tian Computer Vision and Pattern Recognition Vision Transformers (ViTs) have been widely adopted in vision tasks due to their strong transferability. In Federated Learning (FL), where full fine-tuning is communication heavy, Low-Rank Adaptation (LoRA) provides an efficient and communication-friendly way to adapt ViTs. However, existing LoRA-based federated tuning methods overlook latent client structures in real-world settings, limiting shared representation learning and hindering effective adaptation to unseen clients. To address this, we propose HiLoRA, a hierarchical LoRA framework that places adapters at three levels: root, cluster, and leaf, each designed to capture global, subgroup, and client-specific knowledge, respectively. Through cross-tier orthogonality and cascaded optimization, HiLoRA separates update subspaces and aligns each tier with its residual personalized objective. In particular, we develop a LoRA-Subspace Adaptive Clustering mechanism that infers latent client groups via subspace similarity analysis, thereby facilitating knowledge sharing across structurally aligned clients. Theoretically, we establish a tier-wise generalization analysis that supports HiLoRA's design. Experiments on ViT backbones with CIFAR-100 and DomainNet demonstrate consistent improvements in both personalization and generalization. |
| title | HiLoRA: Hierarchical Low-Rank Adaptation for Personalized Federated Learning |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2603.02785 |