AI Paper Atlas

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

Explainable Federated Learning via Global–Local Attribution Alignment

Dawood Wasif, Terrence J Moore, Chang-Tien Lu, Jin-Hee Cho

ICML 2026 · Federated learning and privacy › Personalized and heterogeneous FL · approach: Non-IID federated learning

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 alignment loss (beta 0.2 vs 0) (ablation)xFedAlign · MNISTEDI: 0.0925 -> 0.0841Improved
add alignment loss (beta 0.2 vs 0) (ablation)xFedAlign · MNISTdeletion/insertion AUCNo clear effect
increase alignment weight beta from 0 to 0.4 (ablation)xFedAlign · MNISTEDI: 0.0925 -> 0.0793Improved
add light DP noise (sigma 0.05 vs 0) (ablation)xFedAlign · MNISTEDI: 0.1330 -> 0.0819Improved
raise noise to sigma 0.2 (ablation)xFedAlign · MNISTEDI: 0.0819 -> 0.0852Hurt
increase top-k sparsity from 32 to 512 (ablation)xFedAlign · MNISTEDI: 0.0848 -> 0.0840No clear effect
increase surrogate epochs from 1 to 4 (ablation)xFedAlign · MNISTEDI: 0.0843 -> 0.1161Hurt
use xFedAlign instead of FedAvg-only trainingfederated learning · IMDb non-IIDaccuracy: 0.550 -> 0.702Improved
use xFedAlign instead of FedAvgfederated learning · UCI Credit non-IIDaccuracy: 0.315 -> 0.725Improved
use xFedAlign instead of FedAvgfederated learning · MNIST non-IIDaccuracy: 0.930 -> 0.930No clear effect
use xFedAlign instead of Local-XAIexplanations · IMDb non-IIDEDI: 0.129 -> 0.043Improved
use interpretable-by-design Fed-XAIfederated model · MNIST non-IIDaccuracy: 0.930 -> 0.449Hurt

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