Research map / Adversarial robustness and security / Backdoor and poisoning
FILTER: A Framework for Defending Against Backdoor Attacks in Vertical Federated Learning
Zhanyi Hu, Cen Chen, Yanhao Wang
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 embedding-based filter (ablation) | FILTER · CIFAR-10 ResNet-18 (Optimal Trigger) | ASR: 84.86% -> 0.18% | Improved |
| add embedding-based filter (ablation) | FILTER · CIFAR-10 ResNet-18 (VILLAIN) | ASR: 14.05% -> 10.52% | No clear effect |
| add backdoor unlearning (ablation) | FILTER · CIFAR-10 ResNet-18 (VILLAIN) | ASR: 55.13% -> 10.52% | Improved |
| add backdoor unlearning (ablation) | FILTER · CIFAR-10 ResNet-18 (BadVFL) | ASR: 32.64% -> 28.46% | Improved |
| add backdoor unlearning (ablation) | FILTER · CIFAR-10 ResNet-18 (Mask) | ASR: 28.65% -> 17.41% | Improved |
| add backdoor unlearning (ablation) | FILTER · CIFAR-10 ResNet-18 (VILLAIN) | CDA: 71.06% -> 70.73% | Hurt |
| use Score of Eq. 1 instead of meta-split score (ablation) | FILTER · CIFAR-10 ResNet-18 (BadVFL) | ASR: 83.31% -> 28.46% | Improved |
| use Score of Eq. 1 instead of meta-split score (ablation) | FILTER · CIFAR-10 ResNet-18 (VILLAIN) | ASR: 31.44% -> 10.52% | Improved |
| raise p_init from 0.15 to 0.2 (ablation) | FILTER · CIFAR-10 ResNet-18 (BadVFL) | ASR: 23.25% -> 35.17% | Hurt |
| lower p_init from 0.15 to 0.1 (ablation) | FILTER · CIFAR-10 ResNet-18 (Mask) | CDA: 71.44% -> 68.81% | Hurt |
| set tau to 0.001 instead of 0 (ablation) | FILTER · CIFAR-10 ResNet-18 (Optimal Trigger) | ASR: 94.92% -> 0.18% | Improved |
| raise tau from 0.01 to 0.1 (ablation) | FILTER · CIFAR-10 ResNet-18 (BadVFL) | ASR: 24.36% -> 52.18% | Hurt |
Closest papers by meaning
- Label-Free Backdoor Attacks in Vertical Federated Learning (AAAI 2025 · similarity 0.93)
- VFLAIR: A Research Library and Benchmark for Vertical Federated Learning (ICLR 2024 · similarity 0.91)
- Infighting in the Dark: Multi-Label Backdoor Attack in Federated Learning (CVPR 2025 · similarity 0.89)
- MARS: A Malignity-Aware Backdoor Defense in Federated Learning (NeurIPS 2025 · similarity 0.88)
- Resisting Backdoor Attacks in Federated Learning via Bidirectional Elections and Individual Perspective (AAAI 2024 · similarity 0.88)
- Backdoor Federated Learning by Poisoning Backdoor-Critical Layers (ICLR 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.
