Research map / Federated learning and privacy / Personalized and heterogeneous FL
Bayesian personalized FL
Personalize federated models with Bayesian inference, model reassembly and self-awareness.
35 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2020–2026) · an approach in Personalized and heterogeneous FL, keyword cluster “personalized federated · pfl · personalization”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
14 papers in 2023–24 and 13 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 5.6% to 3.3%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2020 · Personalized Federated Learning with Theoretical Guarantees: A Model-Agnostic Meta-Learning Approach (NeurIPS 2020 · 373 citations)
- 2021 · Ditto: Fair and Robust Federated Learning Through Personalization (ICML 2021 · 216 citations)
- 2022 · Personalized Federated Learning via Variational Bayesian Inference (ICML 2022 · 16 citations)
- 2023 · FedALA: Adaptive Local Aggregation for Personalized Federated Learning (AAAI 2023 · 414 citations)
- 2024 · Fed-QSSL: A Framework for Personalized Federated Learning under Bitwidth and Data Heterogeneity (AAAI 2024 · 8 citations)
- 2025 · FedCALM: Conflict-aware Layer-wise Mitigation for Selective Aggregation in Deeper Personalized Federated Learning (CVPR 2025 · 7 citations)
- 2026 · Breaking the Aggregation Bottleneck in Federated Recommendation: A Personalized Model Merging Approach (AAAI 2026 · 1 citation)
Earliest papers here
- Personalized Federated Learning with Theoretical Guarantees: A Model-Agnostic Meta-Learning Approach (NeurIPS 2020 · 373 citations)
- Ditto: Fair and Robust Federated Learning Through Personalization (ICML 2021 · 216 citations)
- Personalized Federated Learning using Hypernetworks (ICML 2021 · 84 citations)
What has been tried
25 papers here report 276 measured changes (a component added, removed or swapped, compared with a baseline): helped 62%, no clear effect 8%, hurt 25%, mixed 5%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Increase Dirichlet distribution parameter β | CIFAR-10 · Global accuracy | Helped | Fed-QSSL: A Framework for Personalized Federated Learning under Bitwidth and Data Heterogeneity (AAAI 2024 · 8 citations) | |
| add weather decomposition | ODW1T forecasting (Avg.) · MAE | 50.8 -> 45.4 | Helped | Personalized Adapter for Large Meteorology Model on Devices: Towards Weather Foundation Models (NeurIPS 2024 · 6 citations) |
| Set c to 0.2 | Flowers102, C=60, P=30% · accuracy | 58.95 -> 60.20 | Helped | ConFREE: Conflict-free Client Update Aggregation for Personalized Federated Learning (AAAI 2025 · 4 citations) |
| Increase the number of assembled heads from 10 to 50 | Cifar100 · F-ECE | 0.091 -> 0.077 | Helped | Watch Your Head: Assembling Projection Heads to Save the Reliability of Federated Models (AAAI 2024 · 2 citations) |
| Replace random cluster initialization with data-dependent PrivateInit | Reddit, StackOverflow, Amazon Reviews · test NLL | 200 to 20 clustering iterations | Helped | Private and Personalized Frequency Estimation in a Federated Setting (NeurIPS 2024 · 1 citation) |
| Use CIFAR-10 AD containing all 10 classes rather than 8 classes | CIFAR-10 non-IID, medium data · test accuracy | 69.7 ± 0.28 -> 72.7 ± 0.15 | Helped | A Bayesian Approach for Personalized Federated Learning in Heterogeneous Settings (NeurIPS 2024 · 1 citation) |
| Use a separate validation set for router updates instead of newly sampled training batches (Local-MoE-tr / CoMiGS-tr vs CoMiGS) | Multilingual Wikipedia, GPT2-124M · test perplexity | 50.86 (CoMiGS-tr) vs 47.19 (validation set) | Helped | On-Device Collaborative Language Modeling via a Mixture of Generalists and Specialists (ICML 2025 · 0 citations) |
| increase SVGD iterations from 20 to 30 | MNIST, 100 clients · accuracy (%) | 96.59 -> 97.02 | Helped | Personalized Bayesian Federated Learning with Wasserstein Barycenter Aggregation (NeurIPS 2025 · 0 citations) |
| prune first or last two layers in B3 | CIFAR10, class-wise non-iid · accuracy | Helped | FedP3: Federated Personalized and Privacy-friendly Network Pruning under Model Heterogeneity (ICLR 2024 · 0 citations) | |
| remove the conditional mutual information term (λ = 0) | CMNIST, CFMNIST, WaterBird, and PACS · worst-case accuracy | 53.6 -> 20.2; 59.8 -> 18.2; 73.8 -> 64.0; 55.4 -> 46.1 | Hurt | Causally Motivated Personalized Federated Invariant Learning with Shortcut-Averse Information-Theoretic Regularization (ICML 2024) |
What reviewers will expect
Datasets and baselines named in the 6 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- CIFAR10 · 3 of 6 papers
- CIFAR100 · 3 of 6 papers
- Flowers102 · 2 of 6 papers
Baselines
- FedAvg · 3 of 6 papers
- Per-FedAvg · 2 of 6 papers
- FedRep · 2 of 6 papers
- FedCP · 2 of 6 papers
- FedALA · 2 of 6 papers
- FedPAC · 2 of 6 papers
- Local · 2 of 6 papers
- FedProx · 2 of 6 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 27 papers; from OpenAlex and the papers' affiliations).
- University of Texas at Austin · 2 papers
- Central South University · 2 papers
- University of Western Ontario · 2 papers
- Sony · 2 papers
Use it
Download BibTeX of 34 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Personalization with global-local trade-off (48 papers)
- Sparse and collaboration-graph personalization (36 papers)
Elsewhere in the topic
- Non-IID federated learning (198 papers)
- Federated learning with pretrained models (164 papers)
- Federated prompt and LoRA tuning (91 papers)
- Federated semi-supervised and graph learning (84 papers)
- Federated medical and multimodal learning (44 papers)
- Vertical federated learning (38 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).