AI Paper Atlas

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

Divergence-aware Federated Self-Supervised Learning

Weiming Zhuang, Yonggang Wen, Shuai Zhang

ICLR 2022 · Federated learning and privacy › Personalized and heterogeneous FL · approach: Federated semi-supervised and graph learning · 33 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
Use divergence-aware EMA to blend local client parameters with the global modelFedBYOL · CIFAR-10 and CIFAR-100, non-IID linear evaluationtop-1 accuracy: 79.44 ± 0.99 -> 83.34 ± 0.39; 57.51 ± 0.09 -> 61.78 ± 0.14Improved
Use divergence-aware EMA to blend local client parameters with the global modelFedBYOL · CIFAR-10 and CIFAR-100, non-IID semi-supervised evaluationtop-1 accuracy: 70.48 ± 0.30 -> 73.44 ± 0.22; 76.95 ± 0.46 -> 79.49 ± 0.34; 30.21 ± 0.40 -> 33.04 ± 0.23; 47.07 ± 0.14 -> 50.48 ± 0.11Improved
Update both online and target encoders with the global model, removing retained local target-encoder knowledge (ablation)FedBYOL · CIFAR-10, non-IID 2 or 4 classes per clienttop-1 accuracy: 79.44 -> 74.50; 82.82 -> 78.77Hurt
Remove EMA updates of the target encoder (ablation)FedBYOL · CIFAR-10 non-IID linear evaluationtop-1 accuracy: 79.44 -> 50.20Hurt
Remove both EMA and stop-gradient (ablation)FedBYOL · CIFAR-10 non-IID linear evaluationtop-1 accuracy: 79.44 -> 11.97Hurt
Remove EMA and stop-gradient, but update both encoders from the global model (ablation)FedBYOL · CIFAR-10 non-IID linear evaluationtop-1 accuracy: 11.97 -> 68.75Improved
Apply FedEMA only to the predictor instead of to encoder and predictor (ablation)FedEMA · CIFAR-10, linear evaluationtop-1 accuracy: 83.34 -> 81.13Hurt
Apply FedEMA only to the encoder instead of to encoder and predictor (ablation)FedEMA · CIFAR-10, linear evaluationtop-1 accuracy: 83.34 -> 82.39Hurt
Use a constant EMA decay instead of divergence-aware dynamic decay (ablation)FedEMA · CIFAR-10 non-IIDtop-1 accuracyHurt
Use λ > 1, causing the decay μ to reach 1 at the start (ablation)FedEMA · CIFAR datasetstop-1 accuracyHurt
Remove stop-gradient from FedBYOL (ablation)FedBYOL · Non-IID CIFAR-10kNN testing accuracyNo clear effect
Remove the predictor (ablation)FedBYOL and FedSimSiam · Non-IID CIFAR-10kNN testing accuracyHurt

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