AI Paper Atlas

Research map / Federated learning and privacy / Personalized and heterogeneous FL

Identifiable Gated Residual Personalization for Federated Parameter-Efficient Fine-Tuning

Huimin Huang, Wenhan Hu, Gang Yan

ECCV 2026 · Federated learning and privacy › Personalized and heterogeneous FL · approach: Federated prompt and LoRA tuning · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

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 Dynamic Alignment (ablation)FedSDG · CIFAR-100 a=0.1local accuracy: 87.44 -> 88.20Improved
add Dynamic Alignment (ablation)FedSDG · CIFAR-100 a=1.0local accuracy: 77.22 -> 79.24Improved
add Dynamic Alignment (ablation)FedSDG · Tiny-ImageNet a=0.1local accuracy: 79.32 -> 80.63Improved
add Dynamic Alignment (ablation)FedSDG · Tiny-ImageNet a=1.0local accuracy: 68.24 -> 71.54Improved
anchor alignment to frozen backbone instead of moving global branch (ablation)FedSDG · Office-Home dom-0.1local accuracy: 64.64 -> 75.40Improved
anchor alignment to frozen backbone instead of moving global branch (ablation)FedSDG · Tiny-ImageNet a=0.1local accuracy: 80.09 -> 80.63Improved
use learned gates instead of fixed gate m=0.047 (ablation)FedSDG · Tiny-ImageNet a=0.1local accuracy: 76.63 -> 80.63 (+4.0)Improved
use learned gates instead of fixed gate m=0.047 (ablation)FedSDG · Office-Home dom-0.1local accuracy: 70.55 -> 75.40 (+4.8)Improved
raise fixed gate from 0.047 to 1.0 (full personalization) (ablation)FedSDG · Tiny-ImageNet a=0.1local accuracy: 76.63 -> 31.05Hurt
raise fixed gate from 0.047 to 1.0 (full personalization) (ablation)FedSDG · Office-Home dom-0.1local accuracy: 70.55 -> 36.88Hurt
initialize gate logits at -3 instead of 0 (ablation)FedSDG · Tiny-ImageNet a=0.1local accuracy: 52.89 -> 80.63Improved
use projection-level gates instead of coarse gates shared across blocks (ablation)FedSDG · Tiny-ImageNet a=0.1local accuracy: 76.57 -> 80.63Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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