Research map / Federated learning and privacy / Personalized and heterogeneous FL
Local and Global Flatness for Federated Domain Generalization
Hao Yan, Yuhong Guo
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 local flatness regularization (adversarial KL divergence) to ERM (ablation) | FedAvg with additive perturbation · PACS partial | average accuracy (%): 68.8 -> 73.6 | Improved |
| add global flatness regularization to ERM (ablation) | FedAvg with additive perturbation · PACS partial | average accuracy (%): 68.8 -> 75.1 | Improved |
| combine local and global flatness losses (ablation) | FedAvg with additive perturbation · PACS partial | average accuracy (%): 75.1 -> 77.9 | Improved |
| combine local and global flatness losses (ablation) | FedAvg with singular perturbation · PACS partial | average accuracy (%): 78.0 -> 79.3 | Improved |
| use singular instead of additive perturbation (ablation) | FedLGF · Federated DG partial scenario (avg) | test accuracy (%): 78.0 -> 78.4 | Improved |
| use singular instead of additive perturbation (ablation) | FedLGF · Federated DG full scenario (avg) | test accuracy (%): 81.9 -> 81.6 | Hurt |
| apply weight perturbation to feature encoder conv layers as well as classifier (ablation) | FedLGF · federated DG benchmarks | accuracy | Hurt |
| add FedLGF to FedAvg in partial scenario | FedAvg · Federated DG partial (avg) | test accuracy (%): 71.3 -> 78.4 | Improved |
| vary trade-off hyperparameters below 1 (ablation) | FedLGF · PACS partial | average accuracy | No clear effect |
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.
- FedSR — FedSR: A Simple and Effective Domain Generalization Method for Federated Learning (NeurIPS 2022) · better on 1 of 1 shared result
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
- Rethinking the Flat Minima Searching in Federated Learning (ICML 2024 · similarity 0.92)
- Federated Domain Generalization with Data-free On-server Matching Gradient (ICLR 2025 · similarity 0.91)
- FedVSSAM: Mitigating Flatness Incompatibility in Sharpness-Aware Federated Learning (NeurIPS 2026 · similarity 0.90)
- Benchmarking Algorithms for Federated Domain Generalization (ICLR 2024 · similarity 0.90)
- Federated Unsupervised Domain Generalization Using Global and Local Alignment of Gradients (AAAI 2025 · similarity 0.89)
- Efficiently Assemble Normalization Layers and Regularization for Federated Domain Generalization (CVPR 2024 · similarity 0.89)
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.
