AI Paper Atlas

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

Addressing Heterogeneity in Federated Learning via Distributional Transformation

Haolin Yuan, Bo Hui, Yuchen Yang, Philippe Burlina, Neil Zhenqiang Gong, Yinzhi Cao

ECCV 2022 · Federated learning and privacy › Personalized and heterogeneous FL · approach: Non-IID federated learning · 14 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 double-input-channel model instead of single channel (ablation)DisTrans · CH-MNIST (DH 100%)accuracy: 0.889 -> 0.946Improved
use double-input-channel model instead of single channel (ablation)DisTrans · Bird-200 (DH 100%)accuracy: 0.505 -> 0.641Improved
use double-input-channel model instead of single channel (ablation)DisTrans · Bird-200 (DH 0%)accuracy: 0.569 -> 0.556Hurt
use no aggregation instead of NN aggregation at high heterogeneity (ablation)DisTrans · Bird-200 (DH 100%)accuracy: 0.513 -> 0.641Improved
use NN aggregation instead of no aggregation at low heterogeneity (ablation)DisTrans · CH-MNIST (DH 0%)accuracy: 0.868 -> 0.908Improved
use NN aggregation instead of average aggregation (ablation)DisTrans · CIFAR-10 (DH 80%)accuracy: 0.702 -> 0.743Improved
replace FedAvg with DisTransFedAvg · CIFAR-100 (DH 100%)accuracy: 0.524 -> 0.834Improved
replace FedAvg with DisTransFedAvg · CelebA (DH 100%)accuracy: 0.639 -> 0.804Improved
add mixup to FedAvgFedAvg · CH-MNIST (DH 0%)accuracy: 0.891 -> 0.896Improved
add mixup to FedAvgFedAvg · CH-MNIST (DH 100%)accuracy: 0.908 -> 0.901Hurt
add AdvProp to FedAvgFedAvg · CH-MNIST (DH 100%)accuracy: 0.908 -> 0.919Improved
add AdvProp to FedAvgFedAvg · CH-MNIST (DH 0%)accuracy: 0.891 -> 0.879Hurt

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