Research map / Adversarial robustness and security / Backdoor and poisoning
Input-Aware Dynamic Backdoor Attack
Tuan Anh Nguyen, Anh Tran
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 |
|---|---|---|---|
| Use input-conditioned triggers and train with attack, clean, and cross-trigger modes | Input-Aware Dynamic Backdoor · MNIST, CIFAR-10, and GTSRB all-to-all | attack accuracy: 98.47; 93.16; 98.13 | Improved |
| Use input-conditioned triggers and train with attack, clean, and cross-trigger modes | Input-Aware Dynamic Backdoor · MNIST, CIFAR-10, and GTSRB all-to-all | cross-trigger accuracy: 94.34; 89.40; 94.29 | Improved |
| Remove the cross-trigger test (ablation) | Input-Aware Dynamic Backdoor · CIFAR-10 | cross-trigger accuracy: around 10 % | Hurt |
| Remove the cross-trigger test (ablation) | Input-Aware Dynamic Backdoor · CIFAR-10 | Neural Cleanse Anomaly Index: 9.43 | Hurt |
| Add diversity loss to the trigger generator objective (ablation) | Input-Aware Dynamic Backdoor · CIFAR-10 | clean, attack, and cross accuracy: 10% | Improved |
| Add both cross-trigger test and diversity loss (ablation) | Input-Aware Dynamic Backdoor · CIFAR-10 | trigger input-dependence | Improved |
| Increase cross-trigger probability (ablation) | Input-Aware Dynamic Backdoor · CIFAR-10 | cross-trigger accuracy: 80% to 93% | Improved |
| Increase backdoor probability (ablation) | Input-Aware Dynamic Backdoor · CIFAR-10 | attack success rate: near 100% | Improved |
Closest papers by meaning
- Rethinking Backdoor Attacks on Dataset Distillation: A Kernel Method Perspective (ICLR 2024 · similarity 0.90)
- Hidden Trigger Backdoor Attacks (AAAI 2020 · similarity 0.89)
- Versatile Backdoor Attack With Visible, Semantic, Sample-Specific and Compatible Triggers (TPAMI 2025 · similarity 0.89)
- LOTUS: Evasive and Resilient Backdoor Attacks through Sub-Partitioning (CVPR 2024 · similarity 0.88)
- REFINE: Inversion-Free Backdoor Defense via Model Reprogramming (ICLR 2025 · similarity 0.88)
- Demystifying Poisoning Backdoor Attacks from a Statistical Perspective (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.
