Research map / Federated learning and privacy / Personalized and heterogeneous FL
Federated medical and multimodal learning
Harmonize heterogeneous institutions and modalities, including medical imaging domains.
44 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2020–2026) · an approach in Personalized and heterogeneous FL, keyword cluster “domain · label · multi”
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.
11 papers in 2023–24 and 26 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 4.4% to 6.7%.
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 · Federated Patient Hashing (AAAI 2020 · 20 citations)
- 2021 · FedDG: Federated Domain Generalization on Medical Image Segmentation via Episodic Learning in Continuous Frequency Space (CVPR 2021 · 535 citations)
- 2022 · HarmoFL: Harmonizing Local and Global Drifts in Federated Learning on Heterogeneous Medical Images (AAAI 2022 · 157 citations)
- 2023 · Fair Federated Medical Image Segmentation via Client Contribution Estimation (CVPR 2023 · 72 citations)
- 2024 · Think Twice Before Selection: Federated Evidential Active Learning for Medical Image Analysis with Domain Shifts (CVPR 2024 · 35 citations)
- 2025 · FedVCK: Non-IID Robust and Communication-Efficient Federated Learning via Valuable Condensed Knowledge for Medical Image Analysis (AAAI 2025 · 13 citations)
- 2026 · Federated Incomplete Multi-View Clustering with Tensorized Low-Rank Constraint (AAAI 2026 · 2 citations)
Earliest papers here
- Federated Patient Hashing (AAAI 2020 · 20 citations)
- FedDG: Federated Domain Generalization on Medical Image Segmentation via Episodic Learning in Continuous Frequency Space (CVPR 2021 · 535 citations)
- Multi-Institutional Collaborations for Improving Deep Learning-Based Magnetic Resonance Image Reconstruction Using Federated Learning (CVPR 2021 · 176 citations)
What has been tried
27 papers here report 287 measured changes (a component added, removed or swapped, compared with a baseline): helped 80%, no clear effect 5%, hurt 11%, mixed 3%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Add global epistemic uncertainty to global and local aleatoric uncertainty | Fed-ISIC R5 · BMA (%) | 66.70±0.28 -> 68.46±0.37 | Helped | Think Twice Before Selection: Federated Evidential Active Learning for Medical Image Analysis with Domain Shifts (CVPR 2024 · 35 citations) |
| Add IntraGW to FedAvg | SKIN N_S · Dice coefficient (%) | 63.85 ± 2.41 -> 67.45 ± 1.55 | Helped | FedA3I: Annotation Quality-Aware Aggregation for Federated Medical Image Segmentation against Heterogeneous Annotation Noise (AAAI 2024 · 22 citations) |
| Federate the encoder and personalize the decoder rather than federating the decoder | BraTS 2020 validation, Setting 1, client average · mDSC (%) | 38.64 -> 48.58 | Helped | Federated Modality-Specific Encoders and Multimodal Anchors for Personalized Brain Tumor Segmentation (AAAI 2024 · 18 citations) |
| Add CAB without DTD | HAM10000 · accuracy | 68.07 -> 70.29 | Helped | FedCD: Federated Semi-Supervised Learning with Class Awareness Balance via Dual Teachers (AAAI 2024 · 17 citations) |
| Add consensus pre-training | MNIST-USPS, BDGP, Multi-Fashion, and NUSWIDE, M/S = 1:1 · ACC | MNIST-USPS: 91.12 -> 93.38; BDGP: 69.84 -> 91.92; Multi-Fashion: 87.80 -> 90.36; NUSWIDE: 42.16 -> 54.04 | Helped | Bridging Gaps: Federated Multi-View Clustering in Heterogeneous Hybrid Views (NeurIPS 2024 · 17 citations) |
| Use FedST-join instead of FedST-separate (aggregate the generator) | Medical images (CT/MRI) · Dice | 0.9233 -> 0.9221 | No clear effect | FedST: Federated Style Transfer Learning for Non-IID Image Segmentation (AAAI 2024 · 15 citations) |
| Add FCS to DDA | Federated test, average · AUC | 0.793 -> 0.816 | Helped | Point Transformer with Federated Learning for Predicting Breast Cancer HER2 Status from Hematoxylin and Eosin-Stained Whole Slide Images (AAAI 2024 · 15 citations) |
| Add model-guided selection and relational contrastive loss | Path and OrganS, limited and overall communication · accuracy (%) | 72.76 -> 78.52; 71.32 -> 72.65; 73.97 -> 80.36; 71.37 -> 73.23 | Helped | FedVCK: Non-IID Robust and Communication-Efficient Federated Learning via Valuable Condensed Knowledge for Medical Image Analysis (AAAI 2025 · 13 citations) |
| Replace isolated attention with bidirectional self-attention | Open-I and MIMIC-CXR, γ=0.5 and γ=0.1 · AUC | +3.27%; +4.54% | Helped | Incongruent Multimodal Federated Learning for Medical Vision and Language-based Multi-label Disease Detection (AAAI 2025 · 7 citations) |
| add the authenticity module to pseudo bag styles transfer | RCC · AUC (%) | 97.81 -> 98.11 | Helped | HistoFS: Non-IID Histopathologic Whole Slide Image Classification via Federated Style Transfer with RoI-Preserving (CVPR 2025 · 7 citations) |
| Use one-step SFOMVC-TR instead of two-step SVD and K-means | NGs · ACC | 30.40 -> 99.00 | Helped | Scalable Federated One-Step Multi-View Clustering with Tensorized Regularization (AAAI 2025 · 6 citations) |
| add the designed MoE module | Sleep-EDF time-series classification · ACC | 0.8731 -> 0.9041 | Helped | dFLMoE: Decentralized Federated Learning via Mixture of Experts for Medical Data Analysis (CVPR 2025 · 5 citations) |
What reviewers will expect
Datasets and baselines named in the 10 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 10 papers
- MIMIC-CXR · 2 of 10 papers
- WebKB · 2 of 10 papers
Baselines
- FedAvg · 5 of 10 papers
- FedProx · 4 of 10 papers
- MOON · 2 of 10 papers
- FedMVL · 2 of 10 papers
- FedDF · 2 of 10 papers
- FedSM · 2 of 10 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 37 papers; from OpenAlex and the papers' affiliations).
- Beihang University · 2 papers
Use it
Download BibTeX of 29 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
- Non-IID federated learning (198 papers)
- Federated learning with pretrained models (164 papers)
- Federated semi-supervised and graph learning (84 papers)
- Federated prompt and LoRA tuning (91 papers)
- Vertical federated learning (38 papers)
Elsewhere in the topic
- Personalization with global-local trade-off (48 papers)
- Sparse and collaboration-graph personalization (36 papers)
- Bayesian personalized FL (35 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).