Research map / Optimization / Optimizers
FedMDFG: Federated Learning with Multi-Gradient Descent and Fair Guidance
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 a fair descent direction and line-search step size instead of the compared FL updates | Federated learning · CIFAR-10 normal, mutex, and unbalanced | mean test accuracy: 74.6%, 69.1%, and 66.1% | Helped |
| Use FedMDFG rather than the second-best method | Federated learning · CIFAR-10 normal, mutex, and unbalanced | fairness indicator: 33.1%, 46.0%, and 33.2% | Helped |
| Set a smaller tolerable-fair angle θ (ablation) | FedMDFG · CIFAR-10 normal, mutex, and unbalanced | fairness indicator | Helped |
| Add a fair descent direction instead of replacing it with the FedSGD direction (ablation) | FedMDFG · CIFAR-10 and FMNIST mutex | mean test accuracy | Helped |
| Add the step-size line search instead of using η directly (ablation) | FedMDFG · FMNIST and MNIST mutex | mean test accuracy | Mixed |
| Use the default Armijo parameter β=10⁻⁴ rather than β=0 (ablation) | FedMDFG · FMNIST and MNIST mutex | mean test accuracy | No clear effect |
| Use the default Armijo parameter β=10⁻⁴ rather than β=0.5 (ablation) | FedMDFG · FMNIST and MNIST mutex | mean test accuracy | No clear effect |
| Use a smaller fixed η to stabilize updates instead of line search (ablation) | FedMDFG · FMNIST and MNIST mutex | convergence speed | Hurt |
| Reallocate received gradient norms to their average and discard zero-norm updates | FedMDFG · FMNIST normal, client attacks | average test accuracy | Helped |
| Use FedMDFG rather than the compared methods | Federated learning · CIFAR-100 and MNIST mutex | convergence speed | Helped |
Closest papers by meaning
- FedLF: Layer-Wise Fair Federated Learning (AAAI 2024 · similarity 0.92)
- FairTrade: Achieving Pareto-Optimal Trade-Offs between Balanced Accuracy and Fairness in Federated Learning (AAAI 2024 · similarity 0.90)
- FedFACT: A Provable Framework for Controllable Group-Fairness Calibration in Federated Learning (NeurIPS 2025 · similarity 0.90)
- Multi-Dimensional Fair Federated Learning (AAAI 2024 · similarity 0.89)
- Fair Federated Learning under Domain Skew with Local Consistency and Domain Diversity (CVPR 2024 · similarity 0.89)
- Demystifying Local & Global Fairness Trade-offs in Federated Learning Using Partial Information Decomposition (ICLR 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.