Research map / Federated learning and privacy / Personalized and heterogeneous FL
Is Heterogeneity Notorious? Taming Heterogeneity to Handle Test-Time Shift in Federated Learning
Yue Tan, Chen Chen, Weiming Zhuang, Xin Dong, Lingjuan Lyu, Guodong Long
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 changed | Where | Measured | Effect |
|---|---|---|---|
| FedICON vs FedAvg under covariate test-time shift | FedICON · Digit-5 MNIST | test accuracy: 86.18 -> 89.67 | Improved |
| FedICON vs FedAvg under covariate shift | FedICON · Digit-5 USPS | test accuracy: 58.19 -> 72.13 | Improved |
| FedICON vs FedAvg under domain shift | FedICON · Office-10 Caltech | test accuracy: 26.27 -> 33.28 | Improved |
| FedICON vs FedAvg on Office-10 Webcam | FedICON · Office-10 Webcam | test accuracy: 48.21 -> 48.21 | No clear effect |
| add test-time adaptation (Tent) to FedAvg-style training instead of FedICON | Tent · Office-10 Amazon | test accuracy: 44.91 -> 38.25 | Hurt |
| personalized FL (FedRep) instead of generic model | FedRep · Digit-5 MNIST-M | test accuracy: 40.72 -> 19.33 | Hurt |
| FedICON on CIFAR10 attribute shift vs FedAvg, Dir(0.1) | FedICON · CIFAR10 Dir(0.1) | attribute-shift accuracy: 50.09 -> 86.47 | Improved |
| FedICON smallest drop between original and attribute-shift test | FedICON · CIFAR10 Dir(0.1) | accuracy drop Delta: -14.16 -> -3.77 | Improved |
| increase number of clients from 5 to 80 | FedAvg and FedICON · Digit-5 covariate shift | avg test accuracy | Hurt |
| FedICON vs FedAvg across client counts | FedICON · Digit-5 covariate shift | avg test accuracy | Improved |
| increase number of local epochs from 1 to 9 | FedAvg · Digit-5 covariate shift | avg test accuracy | Hurt |
| increase number of local epochs from 1 to 9 | FedICON · Digit-5 covariate shift | avg test accuracy | Improved |
Closest papers by meaning
- FedGPS: Statistical Rectification Against Data Heterogeneity in Federated Learning (NeurIPS 2025 · similarity 0.91)
- Federated Learning with Sample-level Client Drift Mitigation (AAAI 2025 · similarity 0.90)
- Tackling Intertwined Data and Device Heterogeneities in Federated Learning with Unlimited Staleness (AAAI 2025 · similarity 0.89)
- Heterogeneity-Guided Client Sampling: Towards Fast and Efficient Non-IID Federated Learning (NeurIPS 2024 · similarity 0.88)
- A Swiss Army Knife for Heterogeneous Federated Learning: Flexible Coupling via Trace Norm (NeurIPS 2024 · similarity 0.88)
- An Aggregation-Free Federated Learning for Tackling Data Heterogeneity (CVPR 2024 · similarity 0.88)
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.
