Research map / Federated learning and privacy / Privacy attacks and secure learning
FL-WBC: Enhancing Robustness against Model Poisoning Attacks in Federated Learning from a Client Perspective
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 |
|---|---|---|---|
| Perturb small-magnitude local parameter updates with Laplace noise using s=0.4 | FedAvg with FL-WBC · Fashion-MNIST and CIFAR10, IID single-image attack | attack mitigation rounds: 1 communication round | Helped |
| Perturb small-magnitude local parameter updates with Laplace noise using s=0.4 | FedAvg with FL-WBC · Fashion-MNIST and CIFAR10, non-IID single-image attack | attack mitigation rounds: within 5 communication rounds | Helped |
| Use CMA instead of ordinary aggregation | FedAvg · Fashion-MNIST and CIFAR10, single-image attack | attack mitigation rounds: can not mitigate ... within 10 subsequent rounds | Hurt |
| Use CTMA instead of ordinary aggregation | FedAvg · Fashion-MNIST and CIFAR10, single-image attack | attack mitigation rounds: can not be mitigated within 10 subsequent rounds | Hurt |
| Use FL-WBC rather than CDP or LDP | FedAvg · IID single-image attack | attack mitigation rounds: 1 communication round; less than 5% benign accuracy drop | Helped |
| Use FL-WBC rather than CDP or LDP | FedAvg · non-IID single-image attack | attack mitigation rounds: within 5 communication rounds; less than 5% accuracy drop | Helped |
| Combine CMA with FL-WBC | FedAvg · Fashion-MNIST and CIFAR10, IID and non-IID single-image attack | attack mitigation rounds: within 1 communication rounds under IID settings and within 5 communication rounds under non-IID settings | Helped |
| Apply FL-WBC to the multi-image attack setting | FedAvg · IID, 10-image malicious dataset | benign accuracy drop: less than 3% | Helped |
| Apply FL-WBC to the multi-image attack setting | FedAvg · non-IID, 10-image malicious dataset | benign accuracy drop: 10% | Helped |
| Increase the malicious dataset from one image to multiple images | Model poisoning attack · Fashion-MNIST and CIFAR10, IID and non-IID | attack mitigation rounds: can not guarantee mitigating ... within 5 communication rounds | Helped |
Closest papers by meaning
- Model Poisoning Attacks to Federated Learning via Multi-Round Consistency (CVPR 2025 · similarity 0.90)
- Defending Against Sophisticated Poisoning Attacks with RL-based Aggregation in Federated Learning (AAAI 2025 · similarity 0.89)
- Tracing Back the Malicious Clients in Poisoning Attacks to Federated Learning (NeurIPS 2025 · similarity 0.89)
- FedREDefense: Defending against Model Poisoning Attacks for Federated Learning using Model Update Reconstruction Error (ICML 2024 · similarity 0.88)
- Dual Defense: Enhancing Privacy and Mitigating Poisoning Attacks in Federated Learning (NeurIPS 2024 · similarity 0.88)
- RFLPA: A Robust Federated Learning Framework against Poisoning Attacks with Secure Aggregation (NeurIPS 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.