Research map / Federated learning and privacy
Communication-efficient FL: research map
431 papers on Communication-efficient 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 6 approaches. The busiest year so far is 2024.
What it is
Federated learning trains one model across many devices or organisations without collecting their data; each client trains locally and sends updates. This topic analyses the algorithms (FedAvg and successors) and cuts their communication and computation.
Terms you will see on this page:
- Federated learning (FL)
- training across clients that keep their data local.
- FedAvg
- the basic FL algorithm that averages locally trained models.
- Client drift
- local models diverging because clients' data differ.
Start reading with the most cited papers here:
- SCAFFOLD: Stochastic Controlled Averaging for Federated Learning (ICML 2020 · 709 citations)
- SplitFed: When Federated Learning Meets Split Learning (AAAI 2022 · 696 citations)
Within Federated learning and privacy, its share shrank from 19.9% in 2023–24 to 12.7% in 2025–26 (148 → 116 papers at ICML, NeurIPS, ICLR, CVPR and AAAI, the venues with data for all four years).
Explore Communication-efficient 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):
- Communication-efficient heterogeneous FL rising: 13% → 21% of the topic (19 → 24 papers)
- Efficient local training in FL rising: 20% → 27% of the topic (29 → 31 papers)
- Decentralized peer-to-peer training falling: 10% → 4% of the topic (15 → 5 papers)
- Client participation analysis falling: 21% → 17% of the topic (31 → 20 papers)
Papers from 2025–2026 that are already cited most:
- Compressed and distributed least-squares regression: convergence rates with applications to federated learning (ICML 2025 · 104 citations)
- Robust Federated Finetuning of LLMs via Alternating Optimization of LoRA (NeurIPS 2025 · 4 citations)
Approaches and key papers
Federated optimization · 232 papers
Keywords: optimization, convex, fedavg
Approaches in this cluster:
- Client participation analysis (81 papers)
Prove convergence of federated averaging under client heterogeneity and arbitrary participation. - Federated optimization algorithms (76 papers)
Develop faster local-update and composite or minimax optimizers for federated learning. - Local drift correction (75 papers)
Regularize or reaggregate local training to counter heterogeneity and forgetting in federated learning.
Most cited and most cited since 2024:
- SCAFFOLD: Stochastic Controlled Averaging for Federated Learning (ICML 2020 · 709 citations)
- Tackling the Objective Inconsistency Problem in Heterogeneous Federated Optimization (NeurIPS 2020 · 565 citations)
- Compressed and distributed least-squares regression: convergence rates with applications to federated learning (ICML 2025 · 104 citations)
- Communication-Efficient Federated Learning with Accelerated Client Gradient (CVPR 2024 · 38 citations)
Efficient training on edge devices · 199 papers
Keywords: devices, edge, dynamic
Approaches in this cluster:
- Efficient local training in FL (86 papers)
Cut federated training cost with low precision, few rounds and decoupled local learning. - Communication-efficient heterogeneous FL (68 papers)
Reduce communication via dynamic regularization, sparse training and heterogeneity-aware client sampling. - Decentralized peer-to-peer training (45 papers)
Train over decentralized data using asynchronous updates and compressed communication.
Most cited and most cited since 2024:
- SplitFed: When Federated Learning Meets Split Learning (AAAI 2022 · 696 citations)
- HeteroFL: Computation and Communication Efficient Federated Learning for Heterogeneous Clients (ICLR 2021 · 192 citations)
- FedASMU: Efficient Asynchronous Federated Learning with Dynamic Staleness-Aware Model Update (AAAI 2024 · 73 citations)
- Mixed-Precision Quantization for Federated Learning on Resource-Constrained Heterogeneous Devices (CVPR 2024 · 23 citations)
