Research map / Federated learning and privacy / Communication-efficient FL
Federated Dynamic Sparse Training: Computing Less, Communicating Less, Yet Learning Better
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Use dynamic sparse training instead of a fixed random mask | FedDST · non-iid MNIST, cumulative upload caps | best test accuracy: 93.61 -> 96.10; 96.89 -> 97.35; 97.5 -> 97.67; 97.72 -> 97.83 | Helped |
| Use dynamic sparse training instead of a fixed random mask | FedDST · non-iid CIFAR-10, cumulative upload caps | best test accuracy: 33.98 -> 35.41; 41.86 -> 42.27; 45.99 -> 46.72; 48.01 -> 50.67 | Helped |
| Add FedProx to FedDST (ablation) | FedDST · non-iid MNIST | best test accuracy: FedDST: 96.10, 97.35, 97.67, 97.83; FedDST+FedProx: 95.35, 96.97, 97.26, 97.81 | Mixed |
| Add FedProx to FedDST (ablation) | FedDST · non-iid CIFAR-10 | best test accuracy: FedDST: 35.41, 42.27, 46.72, 50.67; FedDST+FedProx: 33.03, 43.18, 46.66, 49.69 | Mixed |
| Vary the mask readjustment ratio α within reasonable values (ablation) | FedDST · non-iid MNIST and CIFAR-10 | accuracy | No clear effect |
| Perform mask readjustments at the end of the local training round (ablation) | FedDST · non-iid MNIST and CIFAR-10 | convergence speed | Helped |
| Vary the target sparsity (ablation) | FedDST · non-iid MNIST and CIFAR-10 | accuracy | No clear effect |
Closest papers by meaning
- SparsyFed: Sparse Adaptive Federated Learning (ICLR 2025 · similarity 0.92)
- An Efficient and Accurate Dynamic Sparse Training Framework Based on Parameter-Freezing (AAAI 2025 · similarity 0.90)
- SpaFL: Communication-Efficient Federated Learning With Sparse Models And Low Computational Overhead (NeurIPS 2024 · similarity 0.89)
- FedQS: Optimizing Gradient and Model Aggregation for Semi-Asynchronous Federated Learning (NeurIPS 2025 · similarity 0.88)
- FedRTS: Federated Robust Pruning via Combinatorial Thompson Sampling (NeurIPS 2025 · similarity 0.88)
- FedMef: Towards Memory-efficient Federated Dynamic Pruning (CVPR 2024 · similarity 0.88)
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.