Research map / Adversarial robustness and security / Backdoor and poisoning
Poisoning with Cerberus: Stealthy and Colluded Backdoor Attack against Federated Learning
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 |
|---|---|---|---|
| jointly fine-tune the trigger, regularize poisoned-model deviation, and diversify malicious updates | CerP · CIFAR-100, Krum | ASR: 89.03 -> 90.76 | Helped |
| jointly fine-tune the trigger, regularize poisoned-model deviation, and diversify malicious updates | CerP · CIFAR-100, Bulyan | ASR: 87.48 -> 99.74 | Helped |
| jointly fine-tune the trigger, regularize poisoned-model deviation, and diversify malicious updates | CerP · CIFAR-100, Foolsgold | ASR: 72.73 -> 91.48 | Helped |
| jointly fine-tune the trigger, regularize poisoned-model deviation, and diversify malicious updates | CerP · LOAN, Krum | ASR: 0 -> 99.98 | Helped |
| jointly fine-tune the trigger, regularize poisoned-model deviation, and diversify malicious updates | CerP · LOAN, Foolsgold | ASR: 0.85 -> 99.97 | Helped |
| fine-tune the trigger rather than use the original trigger (ablation) | CerP on LOAN · LOAN, Krum | ASR: 0 -> 99.98 | Helped |
| add local-model deviation regularization (ablation) | CerP on Fashion-MNIST · Fashion-MNIST, Krum | ASR: 62.46 -> 94.75 | Helped |
| add local-model deviation regularization (ablation) | CerP on LOAN · LOAN, Krum | ASR: 99.98 -> 99.98 | No clear effect |
| add similarity regularization to diversify malicious updates (ablation) | CerP on Fashion-MNIST · Fashion-MNIST, Foolsgold | ASR: 99.80 -> 99.55 | No clear effect |
| add similarity regularization to diversify malicious updates (ablation) | CerP on LOAN · LOAN, Krum | ASR: 99.98 -> 99.98 | No clear effect |
| increase poison local training rounds from 1 to 6 (ablation) | CerP · CIFAR-100 and Fashion-MNIST | ASR | Mixed |
Closest papers by meaning
- Backdoor Federated Learning by Poisoning Backdoor-Critical Layers (ICLR 2024 · similarity 0.90)
- Tracing Back the Malicious Clients in Poisoning Attacks to Federated Learning (NeurIPS 2025 · similarity 0.90)
- Beyond Traditional Threats: A Persistent Backdoor Attack on Federated Learning (AAAI 2024 · similarity 0.88)
- Model Poisoning Attacks to Federated Learning via Multi-Round Consistency (CVPR 2025 · similarity 0.88)
- Resisting Backdoor Attacks in Federated Learning via Bidirectional Elections and Individual Perspective (AAAI 2024 · similarity 0.87)
- Demystifying Poisoning Backdoor Attacks from a Statistical Perspective (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.