Research map / Federated learning and privacy / Personalized and heterogeneous FL
FedShard: Federated Unlearning with Efficiency Fairness and Performance Fairness
Siyuan Wen, Meng Zhang, Yang Yang, Ningning Ding
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 |
|---|---|---|---|
| add cluster-based shard merging A1 to naive sharded FL (ablation) | FedShard · CIFAR-10 (non-IID 0.1, K=32) | training stability: naive sharding collapses -> 56.3% accuracy | Improved |
| add variance-aware round allocation A2 on top of A1 (ablation) | FedShard · CIFAR-10 (non-IID 0.1, K=32) | accuracy (%): 56.3% -> 64.14% | Improved |
| use sharded training instead of FedAvg | FedShard · CIFAR-10 (non-IID 0.1, K=32) | accuracy (%): 62.92% -> 64.14% | Improved |
| use sharded exact unlearning instead of FATS | exact federated unlearning · CIFAR-10 (K=512) | efficiency fairness Me: 1081.73 -> 10.9 | Improved |
| use sharded exact unlearning instead of vanilla retraining | federated unlearning · CIFAR-10 (K=512) | unlearning time (s): 13459.3 -> 3361.1 | Improved |
| use approximate calibration unlearning instead of exact retraining | federated unlearning · CIFAR-10 (K=512) | performance fairness Mp: FedShard 6.08 vs FedEraser 9.78, FedRecovery 19.06 | Hurt |
| use approximate unlearning (RR, FE) instead of retraining-based methods | federated unlearning · CIFAR-10 DPA | DPA precision (%): 0% -> 22.91% (RR), 28.50% (FR) | Hurt |
| use approximate unlearning instead of exact | federated unlearning · CIFAR-10 cascaded leaving | cascaded leaving clients | Hurt |
Closest papers by meaning
- Federated Unlearning with Gradient Descent and Conflict Mitigation (AAAI 2025 · similarity 0.89)
- NoT: Federated Unlearning via Weight Negation (CVPR 2025 · similarity 0.89)
- FairTrade: Achieving Pareto-Optimal Trade-Offs between Balanced Accuracy and Fairness in Federated Learning (AAAI 2024 · similarity 0.89)
- Demystifying Local & Global Fairness Trade-offs in Federated Learning Using Partial Information Decomposition (ICLR 2024 · similarity 0.88)
- Fair Federated Learning under Domain Skew with Local Consistency and Domain Diversity (CVPR 2024 · similarity 0.88)
- Unlearning through Knowledge Overwriting: Reversible Federated Unlearning via Selective Sparse Adapter (CVPR 2025 · 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.
