Research map / Federated learning and privacy / Communication-efficient FL
FedPop: A Bayesian Approach for Personalised Federated Learning
Nikita Yurevich Kotelevskii, Maxime Vono, Alain Durmus, Eric Moulines
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 |
|---|---|---|---|
| FedSOUL vs best baseline FedRep | FedSOUL · CIFAR-10 (100,2) | accuracy: 87.70 -> 91.12 | Improved |
| FedSOUL vs best baseline FedRep | FedSOUL · CIFAR-10 (100,5) | accuracy: 75.68 -> 79.48 | Improved |
| FedSOUL vs best baseline FedRep | FedSOUL · CIFAR-100 (100,20) | accuracy: 56.10 -> 59.73 | Improved |
| FedSOUL vs best baseline DITTO | FedSOUL · CIFAR-100 (100,5) | accuracy: 79.15 -> 79.56 | Improved |
| FedSOUL vs local learning only | FedSOUL · CIFAR-10 (100,2) | accuracy: 89.79 -> 91.12 | Improved |
| add fine-tuning to FedAvg | FedAvg · CIFAR-10 (100,2) | accuracy: 42.65 -> 85.63 | Improved |
| add fine-tuning to FedAvg | FedAvg · CIFAR-100 (100,20) | accuracy: 31.97 -> 56.19 | Improved |
| non-personalised SCAFFOLD vs FedAvg | SCAFFOLD · CIFAR-10 (100,2) | accuracy: 42.65 -> 37.72 | Hurt |
| use common prior with small synthetic datasets (FedSOUL vs FedRep/FedAvg) | FedSOUL · synthetic regression, small data | principal angle distance and l2 norm | Improved |
| FedSOUL vs FedRep with 10%, 50%, 90% small clients | FedSOUL · CIFAR-10/100 small data | average accuracy | Improved |
| run FedPop without warm-start, needing M=50 (ablation) | FedSOUL · synthetic | performance: M=50 | No clear effect |
| compare FedPop with non-federated stochastic approximation, T=100 | FedPop · synthetic | relative l2 error: less than 10^-3 | Improved |
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.
- DITTO — Ditto: Fair and Robust Federated Learning Through Personalization (ICML 2021) · better on 4 of 4 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
- A Bayesian Approach for Personalized Federated Learning in Heterogeneous Settings (NeurIPS 2024 · similarity 0.90)
- Harnessing Heterogeneous Statistical Strength for Personalized Federated Learning via Hierarchical Bayesian Inference (ICML 2025 · similarity 0.89)
- FedAS: Bridging Inconsistency in Personalized Federated Learning (CVPR 2024 · similarity 0.89)
- FedPop: Federated Population-based Hyperparameter Tuning (AAAI 2025 · similarity 0.88)
- FedGPS: Statistical Rectification Against Data Heterogeneity in Federated Learning (NeurIPS 2025 · similarity 0.88)
- Understanding the Statistical Accuracy-Communication Trade-off in Personalized Federated Learning with Minimax Guarantees (ICML 2025 · similarity 0.87)
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.
