AI Paper Atlas

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

FedUSD: Unbiased Synthetic Data for Federated Learning

Weiying Xie, Chenhe Hao, Haozhi Shi, Jitao Ma, Daixun Li, Jiazhe Li, Hengyi Wang, Leyuan Fang, Yunsong Li

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
use FedUSD instead of FedAFaggregation-free FL, ConvNet · CIFAR-10 (alpha=0.1)Top-1 accuracy (%): 67.38 -> 74.73Improved
use FedUSD instead of FedAFConvNet · CIFAR-100 (alpha=0.01)Top-1 accuracy (%): 34.53 -> 44.58Improved
use FedUSD instead of FedAFConvNet · SVHN (alpha=0.1)Top-1 accuracy (%): 84.32 -> 89.96Improved
use FedUSD instead of FedAFViT-Tiny-16 · CIFAR-10 (alpha=0.1)Top-1 accuracy (%): 47.81 -> 57.07Improved
add SEA module to baseline (ablation)FedUSD baseline, ConvNet · CIFAR-10 (alpha=0.1)Top-1 accuracy (%): 62.15 -> 72.50Improved
add USD loss to baseline (ablation)FedUSD baseline, ConvNet · CIFAR-10 (alpha=0.1)Top-1 accuracy (%): 62.15 -> 69.61Improved
add USD loss to baseline with SEA (full FedUSD) (ablation)FedUSD, ConvNet · CIFAR-10 (alpha=0.1)Top-1 accuracy (%): 72.50 -> 74.73Improved
add USD loss and SEA to baseline (ablation)FedUSD, ConvNet · CIFAR-100 (alpha=0.01)Top-1 accuracy (%): 28.61 -> 44.58Improved
increase IPC from 1 to 10 (ablation)FedUSD, ConvNet · CIFAR-10 (alpha=0.1)Top-1 accuracy (%): 48.14 -> 74.73Improved
increase IPC from 10 to 50 (ablation)FedUSD, ConvNet · CIFAR-10 (alpha=0.1)Top-1 accuracy (%): 74.73 -> 81.14Improved
use FedUSD instead of FedAvg for communicationConvNet, alpha=0.01 · CIFAR-10communication cost: 256.0MB -> 35.2MBImproved
use FedUSD instead of FedDM for communicationConvNet, alpha=0.01 · CIFAR-10communication cost: 35.2MB -> 35.2MBNo clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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