Research map / Federated learning and privacy / Privacy attacks and secure learning
Batman: Benign Knowledge Alignment Through Malicious Null Space in Federated Backdoor Attack
Wenwen He, Wenke Huang, Yiyang Fang, Wenjie Qu, Jiaheng Zhang, Mang Ye
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 BeKA (benign knowledge alignment) to MaSC-only (ablation) | Batman · CINIC (FLTrust) | ASR: 76.91 -> 87.76 | Improved |
| add BeKA to MaSC-only (ablation) | Batman · Fashion-MNIST (FLTrust) | ASR: 78.10 -> 99.74 | Improved |
| add MaSC (malicious null space) to BeKA-only (ablation) | Batman · CINIC (FLTrust) | ASR: 67.66 -> 87.76 | Improved |
| add MaSC to BeKA-only (ablation) | Batman · Fashion-MNIST (FLTrust) | ASR: 60.60 -> 99.74 | Improved |
| add MaSC to BeKA-only (ablation) | Batman · CINIC (DnC) | ASR: 96.09 -> 98.41 | Improved |
| add BeKA to MaSC-only (ablation) | Batman · Fashion-MNIST (DnC) | ASR: 99.50 -> 99.59 | No clear effect |
| use clean-local and global alignment strength [20,15] vs low alignment (ablation) | Batman · CIFAR-10/CINIC hyperparameters | AVG: 87.76 (best) | Improved |
| use rank r=4 in truncated SVD (ablation) | Batman · CIFAR-10 and CINIC | AVG | Mixed |
| use Batman instead of BadNet against FLTrust | Batman · CIFAR-10 (FLTrust) | ASR: 66.11 -> 85.52 | Improved |
| use Batman instead of BadNet against AlignIns | Batman · CINIC (AlignIns) | ASR: 24.59 -> 99.08 | Improved |
| use Batman instead of BadNet against FedAvg | Batman · CIFAR-10 (FedAvg) | ASR: 95.06 -> 97.02 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Neurotoxin — Neurotoxin: Durable Backdoors in Federated Learning (ICML 2022) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- MARS: A Malignity-Aware Backdoor Defense in Federated Learning (NeurIPS 2025 · similarity 0.89)
- Detecting Backdoor Attacks in Federated Learning via Direction Alignment Inspection (CVPR 2025 · similarity 0.89)
- SPMC: Self-Purifying Federated Backdoor Defense via Margin Contribution (ICML 2025 · similarity 0.89)
- Beyond Traditional Threats: A Persistent Backdoor Attack on Federated Learning (AAAI 2024 · similarity 0.88)
- Parameter Disparities Dissection for Backdoor Defense in Heterogeneous Federated Learning (NeurIPS 2024 · similarity 0.88)
- Resisting Backdoor Attacks in Federated Learning via Bidirectional Elections and Individual Perspective (AAAI 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.
