Research map / Federated learning and privacy / Personalized and heterogeneous FL
The Gaussian-Head OFL Family: One-Shot Federated Learning from Client Global Statistics
Fabio Turazza, Marco Picone, Marco Mamei
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 |
|---|---|---|---|
| model class-specific covariances (QDA instead of LDA) (ablation) | GH-LDA · CIFAR-100-C | accuracy (%): 37.6% -> 64.3% | Improved |
| use Fisher-space trainable head (FisherMix) instead of LDA (ablation) | GH-LDA · CIFAR-100-C | accuracy (%): 37.6% -> 40.1% | Improved |
| replace full QDA with LDA on clean CIFAR-10 (ablation) | GH-QDA_full · CIFAR-10 | accuracy (%): 84.40 -> 86.05 | Improved |
| use Places365 instead of ImageNet pretraining for ResNet-18 (QDA vs LDA on CIFAR-100-C) (ablation) | GH-LDA · CIFAR-100-C | accuracy (%): 37.6% -> 25.99 | Hurt |
| use ResNet50 backbone instead of VGG16 (ablation) | GH-LDA · CIFAR-10 | accuracy (%): 81.39 -> 91.26 | Improved |
| use EfficientNet-B0 backbone instead of ResNet18 (ablation) | GH-LDA · CIFAR-10 | accuracy (%): 86.17 -> 90.01 | Improved |
| use ResNet50 instead of ResNet18 with QDA (ablation) | GH-QDA_full · CIFAR-10 | accuracy (%): 84.76 -> 86.68 | Improved |
| use full QDA instead of LDA in RP space (ablation) | LDA · AG_NEWS | accuracy (%): 88.93 -> 68.20 | Hurt |
| increase DLR-QDA rank from 4 to 16 (ablation) | DLR-QDA · AG_NEWS | accuracy (%): 67.08 -> 38.32 | Hurt |
| increase DLR-QDA rank from 4 to 16 (ablation) | DLR-QDA · CIFAR-100 (RP space) | accuracy (%): 49.99 -> 42.01 | Hurt |
| increase DLR-QDA rank from 4 to 16 (ablation) | DLR-QDA · DBPEDIA-14 | accuracy (%): 92.32 -> 92.19 | No clear effect |
| use GH-LDA one-shot instead of FedAvg 1 round | FedAvg · CIFAR-10 (a=0.1) | accuracy (%): 24.53 -> 86.05 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- Co-Boosting — Co-Boosting++: Coupled Optimization of Data and Ensemble for One-Shot Federated Learning (TPAMI 2026) · better on 3 of 3 shared results
- DENSE — DENSE: Data-Free One-Shot Federated Learning (NeurIPS 2022) · 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
- FedBEns: One-Shot Federated Learning based on Bayesian Ensemble (ICML 2025 · similarity 0.89)
- FedLPA: One-shot Federated Learning with Layer-Wise Posterior Aggregation (NeurIPS 2024 · similarity 0.89)
- FuseFL: One-Shot Federated Learning through the Lens of Causality with Progressive Model Fusion (NeurIPS 2024 · similarity 0.88)
- Capture Global Feature Statistics for One-Shot Federated Learning (AAAI 2025 · similarity 0.88)
- Revisiting Ensembling in One-Shot Federated Learning (NeurIPS 2024 · similarity 0.88)
- One-Shot Heterogeneous Federated Learning with Local Model-Guided Diffusion Models (ICML 2025 · 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.
