Research map / Federated learning and privacy / Communication-efficient FL
Enhancing Communication Compression via Discrepancy-aware Calibration for Federated Learning
Zhiyi Wan, Yijia Chi, Liang Li, Wanrou Du, Miao Pan, Xiaoqi Qin
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| replace magnitude-based selection with discrepancy-aware selection | Top-k · CIFAR-10 (ViT-tiny), ratio 0.01 | final test accuracy: 21.03 -> 29.62 | Improved |
| replace magnitude-based selection with discrepancy-aware selection | Top-k · CIFAR-100 (ViT-small), ratio 0.01 | final test accuracy: 9.29 -> 13.29 | Improved |
| replace magnitude-based selection with discrepancy-aware selection | Top-k · CIFAR-100 (ResNet-18), ratio 0.01 | final test accuracy: 10.76 -> 15.58 | Improved |
| replace magnitude-based selection with discrepancy-aware selection | Top-k · Fashion-MNIST (AlexNet), ratio 0.01 | final test accuracy: 63.31 -> 67.55 | Improved |
| replace magnitude-based selection with discrepancy-aware selection | Top-k · CIFAR-10 (ViT-tiny), ratio 0.6 | final test accuracy: 38.57 -> 41.21 | Improved |
| replace magnitude-based selection with discrepancy-aware selection | ATOMO · CIFAR-100 (ViT-base), rank 1 | final test accuracy: 17.08 -> 20.17 | Improved |
| replace magnitude-based selection with discrepancy-aware selection | ATOMO · CIFAR-10 (ViT-tiny), rank 8 | final test accuracy: 42.44 -> 42.68 | No clear effect |
| replace magnitude-based selection with discrepancy-aware selection | ATOMO · CIFAR-10 (ViT-small), rank 2 | final test accuracy: 37.01 -> 40.10 | Improved |
| add discrepancy-aware selection to reduce rounds to 80% target | Top-k · CIFAR-10 (ViT-tiny), ratio 0.01 | communication rounds: 128 -> 82 (1.56x) | Improved |
| add discrepancy-aware selection to reduce rounds to 80% target | Top-k · FMNIST (AlexNet), ratio 0.2 | communication rounds: 83 -> 83 (1.00x) | No clear effect |
| use random calibration samples instead of first samples (ablation) | discrepancy-aware Top-k · Fashion-MNIST (AlexNet) | final test accuracy: 72.49 -> 73.95 | Improved |
| use random calibration samples instead of last samples (ablation) | discrepancy-aware Top-k · CIFAR-10 (ViT-tiny) | final test accuracy: 42.91 -> 43.23 | Improved |
Closest papers by meaning
- Stochastic Controlled Averaging for Federated Learning with Communication Compression (ICLR 2024 · similarity 0.89)
- An Efficient and Accurate Dynamic Sparse Training Framework Based on Parameter-Freezing (AAAI 2025 · similarity 0.88)
- Achieving Dimension-Free Communication in Federated Learning via Zeroth-Order Optimization (ICLR 2025 · similarity 0.88)
- Achieving Lossless Gradient Sparsification via Mapping to Alternative Space in Federated Learning (ICML 2024 · similarity 0.88)
- FedSMU: Communication-Efficient and Generalization-Enhanced Federated Learning through Symbolic Model Updates (ICML 2025 · similarity 0.87)
- z-SignFedAvg: A Unified Stochastic Sign-Based Compression for Federated Learning (AAAI 2024 · similarity 0.87)
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.
