Research map / Adversarial robustness and security / Backdoor and poisoning
Backdoor Attack with Imperceptible Input and Latent Modification
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 |
|---|---|---|---|
| add sliced-Wasserstein latent regularization R_phi to the backdoor objective (ablation) | WB / LIRA-style backdoor · MNIST and CIFAR10 activation clustering | latent cluster separability | Helped |
| add latent Wasserstein regularization (WB vs LIRA) | LIRA · CIFAR10 all-to-one | attack success rate: 1.00 -> 0.99 | Hurt |
| add latent Wasserstein regularization (WB vs LIRA) | LIRA · MNIST all-to-all | attack success rate: 0.99 -> 0.96 | Hurt |
| use WB instead of Adversarial Embedding against activation clustering | DenseNet CIFAR10 · adjusted Rand Index all-to-one | adjusted Rand Index: 0.1820 -> 0.0382 | Helped |
| use WB instead of Adversarial Embedding against activation clustering | VGG GTSRB · adjusted Rand Index all-to-one | adjusted Rand Index: 0.6420 -> 0.1010 | Helped |
| use discriminant-direction sliced Wasserstein (DSWD) instead of random-direction SWD (ablation) | latent distance estimate · MNIST/CIFAR10 latent space | estimate variance and compute | Helped |
Closest papers by meaning
- BAN: Detecting Backdoors Activated by Adversarial Neuron Noise (NeurIPS 2024 · similarity 0.89)
- Attention-Imperceptible Backdoor Attacks on Vision Transformers (AAAI 2025 · similarity 0.88)
- BaDExpert: Extracting Backdoor Functionality for Accurate Backdoor Input Detection (ICLR 2024 · similarity 0.88)
- A Dual Stealthy Backdoor: From Both Spatial and Frequency Perspectives (AAAI 2024 · similarity 0.88)
- REFINE: Inversion-Free Backdoor Defense via Model Reprogramming (ICLR 2025 · similarity 0.88)
- Adversarial-Inspired Backdoor Defense via Bridging Backdoor and Adversarial Attacks (AAAI 2025 · 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.