Research map / Federated learning and privacy / Personalized and heterogeneous FL
Domain-Skewed Federated Learning with Feature Decoupling and Calibration
Huan Wang, Jun Shen, Jun Yan, Guansong Pang
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 |
|---|---|---|---|
| Add domain-aware aggregation (ablation) | F2DC · PACS | AVG accuracy: 75.33 -> 76.47 | Improved |
| Add domain-aware aggregation (ablation) | F2DC · Office-Caltech | AVG accuracy: 65.29 -> 66.82 | Improved |
| Plug DFD and DFC into FedAvg | FedAvg · PACS | AVG accuracy: +8.94 | Improved |
| Plug DFD and DFC into FedHEAL | FedHEAL · PACS | AVG accuracy: +1.72 | Improved |
| Plug DFD and DFC into FDSE | FDSE · PACS | AVG accuracy: +1.66 | Improved |
| Use corrected features f* vs domain-related features f- (ablation) | F2DC · PACS | AVG accuracy: 57.87 -> 73.49 | Improved |
| Combine f+ with corrected f* instead of f+ alone (ablation) | F2DC · PACS | AVG accuracy: 75.13 -> 76.47 | Improved |
| Contrastive alignment to global representation (MOON, FedRCL) | FedAvg · PACS | AVG accuracy: 66.39 -> 62.64 | Hurt |
| Use very small temperature tau=0.02 to force all signals into f+ (ablation) | F2DC · Office-Caltech, PACS | AVG accuracy | Hurt |
| Set mask smoothness sigma too small or too large (ablation) | F2DC · PACS | AVG accuracy | Hurt |
| Extremely large lambda1 or lambda2 (ablation) | F2DC · Office-Caltech, PACS | AVG accuracy | Hurt |
| Add decoupler and corrector computation | F2DC · PACS | training time per round: 176.94s (FedAvg) vs 180.67s | Mixed |
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.
- FedSA — FedSA: A Unified Representation Learning via Semantic Anchors for Prototype-based Federated Learning (AAAI 2025) · better on 3 of 3 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
- Federated Learning with Domain Shift Eraser (CVPR 2025 · similarity 0.91)
- FedWon: Triumphing Multi-domain Federated Learning Without Normalization (ICLR 2024 · similarity 0.90)
- A Simple Data Augmentation for Feature Distribution Skewed Federated Learning (CVPR 2025 · similarity 0.89)
- Fair Federated Learning under Domain Skew with Local Consistency and Domain Diversity (CVPR 2024 · similarity 0.89)
- Federated Domain Generalization with Data-free On-server Matching Gradient (ICLR 2025 · similarity 0.88)
- Federated Graph Learning under Domain Shift with Generalizable Prototypes (AAAI 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.
