AI Paper Atlas

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

SFedPO: Streaming Federated Learning with a Prediction Oracle under Temporal Shifts

Jinrui Zhou, Haotian Xu, Xichong zhang, He Sun, Mingjun Xiao

ICML 2026 · Federated learning and privacy › Personalized and heterogeneous FL · approach: Federated learning with pretrained models

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 SFedPO sampling and aggregation to FedAvgFedAvg · CIFAR-10 (partial access)test accuracy (%): +4.01Improved
add SFedPO sampling and aggregation to FedEXPFedEXP · CIFAR-10 (partial access)test accuracy (%): +7.76Improved
add SFedPO sampling and aggregation to FedAvgFedAvg · Fashion-MNIST (partial access)test accuracy (%): +0.13Improved
add SFedPO sampling and aggregation to FedProxFedProx · CIFAR-10 (partial access)test accuracy (%): +4.58Improved
add SFedPO sampling and aggregation to FedNTDFedNTD · CINIC-10 (partial access)test accuracy (%): +0.80Improved
add the Distribution-guided Data Sampling module (vs uniform sampling across latent states) (ablation)SFedPO · CIFAR-10 (partial access)test accuracy (%)Improved
add the Shift-aware Aggregation Weights module (vs uniform aggregation) (ablation)SFedPO · CIFAR-10 / CINIC-10test accuracy (%)Improved
add oracle prediction noise eps from 0.00 to 0.10SFedPO · CIFAR-10 (full access)test accuracy (%): 67.45 -> 66.95No clear effect
add oracle prediction noise eps from 0.00 to 0.10SFedPO · CIFAR-10 (partial access)test accuracy (%): 63.52 -> 61.90Hurt
use SFedPO instead of DRSRcoreset/data selection · CIFAR-10 (partial access)test accuracy (%): 48.42 -> 63.53Improved
use SFedPO instead of Flash with VGG16FL method · CIFAR-10 (full access, VGG16)test accuracy (%): 27.45 -> 54.64Improved

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