AI Paper Atlas

Research map / Federated learning and privacy / Differential privacy

Taming Noise-Induced Prototype Degradation for Privacy-Preserving Personalized Federated Fine-Tuning

Yuhua Wang, Qinnan Zhang, Xiaodong Li, Huan Zhang, Yifan Sun, Wangjie Qiu, Hainan Zhang, Yongxin Tong, Zhiming Zheng

CVPR 2026 · Federated learning and privacy › Differential privacy · approach: Differentially private federated learning · 0 citations (OpenAlex)

Read the paper

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add VPP (variance-adaptive prototype perturbation) to IGPP (ablation)FedProto · Office-Caltechaverage accuracy (AVG): 90.82 -> 93.06Improved
add DCR (distillation-guided clipping regularization) to IGPP (ablation)FedProto · Office-Caltechaverage accuracy (AVG): 90.82 -> 92.89Improved
add VPP to IGPP (ablation)FedProto · PACSaverage accuracy (AVG): 90.58 -> 91.89Improved
add DCR to IGPP (ablation)FedProto · PACSaverage accuracy (AVG): 90.58 -> 91.91Improved
add DCR on top of VPP (full VPDR) (ablation)FedProto · PACSaverage accuracy (AVG): 91.89 -> 92.71Improved
add DCR on top of VPP (full VPDR) (ablation)FedProto · Office-Caltechaverage accuracy (AVG): 93.06 -> 93.36Improved
replace IGPP with VPDRFedPCL · PACS (domain skew, eps=1)average accuracy (AVG): 79.73 -> 86.91Improved
replace IGPP with VPDRFedProto · Digits (domain skew, eps=1)average accuracy (AVG): 94.36 -> 96.05Improved
replace IGPP with VPDR under extreme label skew (alpha=0.1)FPL · CIFAR-10 label skewaverage accuracy (AVG): 28.42 -> 37.18Improved
replace IGPP with VPDR under extreme label skew (alpha=0.1)FedProto · CIFAR-10 label skewaverage accuracy (AVG): 42.76 -> 44.14Improved
replace IGPP with VPDR at alpha=1FedPLVM · CIFAR-10 label skewaverage accuracy (AVG): 87.21 -> 82.76Hurt
replace IGPP with VPDR at near-balanced labels (alpha=10)FedProto · CIFAR-10 label skewaverage accuracy (AVG): 96.30 -> 96.55Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy