Research map / Federated learning and privacy
Personalized and heterogeneous FL: research map
860 papers on Personalized and heterogeneous FL in Federated learning and privacy, from ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR (2016–2026), grouped into 2 clusters and 9 approaches. The busiest year so far is 2026.
What it is
Clients in federated learning have very different data (non-IID). This topic handles that heterogeneity and personalises the model to each client while still sharing what is common, including with pretrained models, prompts and LoRA.
Terms you will see on this page:
- Non-IID data
- client datasets that are not drawn from the same distribution.
- Personalized FL
- learning a tailored model for each client.
- Vertical FL
- clients that hold different features of the same users.
Start reading with the most cited papers here:
- Model-Contrastive Federated Learning (CVPR 2021 · 1,518 citations)
- FedProto: Federated Prototype Learning across Heterogeneous Clients (AAAI 2022 · 751 citations)
Within Federated learning and privacy, its share grew from 33.5% in 2023–24 to 42.6% in 2025–26 (249 → 389 papers at ICML, NeurIPS, ICLR, CVPR and AAAI, the venues with data for all four years).
Explore Personalized and heterogeneous FL in the interactive map
Working on something in this topic? Describe your idea in scime atlas to see which approach it falls under, the closest papers by meaning and how crowded the spot has become.
What changed recently
Approaches whose share of the topic moved most between 2023–24 and 2025–26 (ICML, NeurIPS, ICLR, CVPR and AAAI):
- Federated prompt and LoRA tuning rising: 7% → 18% of the topic (17 → 69 papers)
- Federated semi-supervised and graph learning rising: 8% → 14% of the topic (20 → 54 papers)
- Federated medical and multimodal learning rising: 4% → 7% of the topic (11 → 26 papers)
- Federated learning with pretrained models falling: 27% → 19% of the topic (67 → 75 papers)
- Vertical federated learning falling: 7% → 3% of the topic (18 → 10 papers)
Papers from 2025–2026 that are already cited most:
- Patient-Level Anatomy Meets Scanning-Level Physics: Personalized Federated Low-Dose CT Denoising Empowered by Large Language Model (CVPR 2025 · 18 citations)
- Personalized Federated Collaborative Filtering: A Variational AutoEncoder Approach (AAAI 2025 · 15 citations)
- FedCALM: Conflict-aware Layer-wise Mitigation for Selective Aggregation in Deeper Personalized Federated Learning (CVPR 2025 · 7 citations)
Approaches and key papers
Heterogeneous federated learning · 725 papers
Keywords: domain, label, multi
Approaches in this cluster:
- Non-IID federated learning (224 papers)
Counter data heterogeneity with drift decoupling, distillation and re-labeling. - Federated learning with pretrained models (188 papers)
Handle client imbalance and long-tailed data with pretrained and CLIP models and consistency. - Federated semi-supervised and graph learning (96 papers)
Learn from unlabeled or graph data across clients with pseudo-labeling. - Federated prompt and LoRA tuning (109 papers)
Adapt foundation models federatedly with prompts and low-rank adapters. - Federated medical and multimodal learning (64 papers)
Harmonize heterogeneous institutions and modalities, including medical imaging domains. - Vertical federated learning (44 papers)
Train across parties holding different features with privacy and communication efficiency.
Most cited and most cited since 2024:
- Model-Contrastive Federated Learning (CVPR 2021 · 1,518 citations)
- FedProto: Federated Prototype Learning across Heterogeneous Clients (AAAI 2022 · 751 citations)
- FedTGP: Trainable Global Prototypes with Adaptive-Margin-Enhanced Contrastive Learning for Data and Model Heterogeneity in Federated Learning (AAAI 2024 · 114 citations)
- FedDAT: An Approach for Foundation Model Finetuning in Multi-Modal Heterogeneous Federated Learning (AAAI 2024 · 82 citations)
Personalized federated learning · 135 papers
Keywords: personalized federated, pfl, personalization
Approaches in this cluster:
- Personalization with global-local trade-off (55 papers)
Balance global and personalized federated models with sharpness, prototypes and Gaussian processes. - Sparse and collaboration-graph personalization (42 papers)
Adapt models to clients via sparse adaptation, latent constraints and inferred collaboration graphs. - Bayesian personalized FL (38 papers)
Personalize federated models with Bayesian inference, model reassembly and self-awareness.
Most cited and most cited since 2024:
- FedALA: Adaptive Local Aggregation for Personalized Federated Learning (AAAI 2023 · 414 citations)
- Personalized Federated Learning with Theoretical Guarantees: A Model-Agnostic Meta-Learning Approach (NeurIPS 2020 · 373 citations)
- FedAS: Bridging Inconsistency in Personalized Federated Learning (CVPR 2024 · 99 citations)
- FedSelect: Personalized Federated Learning with Customized Selection of Parameters for Fine-Tuning (CVPR 2024 · 38 citations)
