Research map / Adversarial robustness and security / Backdoor and poisoning
SPD: Shallow Backdoor Protecting Deep Backdoor Against Backdoor Detection
Shunjie Yuan, Xinghua Li, Xuelin Cao, Haiyan Zhang, Mengyao Zhu, Robert H. Deng
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 shallow backdoor (ablation) | SPD · CIFAR-10 | MM-BD detection success rate (%): 90.0 -> 0.0 | Improved |
| add shallow backdoor (ablation) | SPD · CIFAR-10 | NC detection success rate (%): 100.0 -> 0.0 | Improved |
| add shallow backdoor (ablation) | SPD · CIFAR-10 | BTI-DBF detection success rate (%): 100.0 -> 0.0 | Improved |
| add shallow backdoor (ablation) | SPD · GTSRB | MM-BD detection success rate (%): 100.0 -> 10.0 | Improved |
| increase alpha from 5.0 to 100.0 (ablation) | SPD · CIFAR-10 | PSNR: 50.75 -> 74.52 | Improved |
| increase alpha from 5.0 to 50.0 (ablation) | SPD · CIFAR-10 | attack success rate (%): 99.91 -> 11.31 | Hurt |
| increase alpha from 5.0 to 50.0 (ablation) | SPD · CIFAR-10 | clean accuracy (%): 94.38 -> 88.96 | Hurt |
| increase alpha from 10.0 to 20.0 (ablation) | SPD · CIFAR-10 | MM-BD detection success rate (%): 10% -> 20% | Hurt |
| increase alpha from 10.0 to 20.0 (ablation) | SPD · CIFAR-10 | BTI-DBF detection success rate (%): 0% -> 10% | Hurt |
| increase alpha from 5.0 to 100.0 (ablation) | SPD · GTSRB | PSNR: 37.21 -> 46.22 | Improved |
| use SPD instead of Ftrojann | SPD · CIFAR-10 | PSNR: 40.47 -> 50.75 | Improved |
| use SPD instead of Ftrojann | SPD · CIFAR-10 | clean accuracy (%): 93.56 -> 94.38 | Improved |
Closest papers by meaning
- A Dual Stealthy Backdoor: From Both Spatial and Frequency Perspectives (AAAI 2024 · similarity 0.88)
- WaveAttack: Asymmetric Frequency Obfuscation-based Backdoor Attacks Against Deep Neural Networks (NeurIPS 2024 · similarity 0.87)
- BAN: Detecting Backdoors Activated by Adversarial Neuron Noise (NeurIPS 2024 · similarity 0.86)
- BackdoorDM: A Comprehensive Benchmark for Backdoor Learning on Diffusion Model (NeurIPS 2025 · similarity 0.86)
- Attention-Imperceptible Backdoor Attacks on Vision Transformers (AAAI 2025 · similarity 0.86)
- Backdoor Defense via Test-Time Detecting and Repairing (CVPR 2024 · similarity 0.86)
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.
