Research map / Adversarial robustness and security / Backdoor and poisoning
Augmented Neural Fine-tuning for Efficient Backdoor Purification
Nazmul Karim, Abdullah Al Arafat, Umar Khalid, Zhishan Guo, Nazanin Rahnavard
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 mask regularizer to NFT (ablation) | NFT · CIFAR-10 LIRA, 10 samples | ACC: 74.58 -> 82.72 | Improved |
| add MixUp augmentation to NFT (ablation) | NFT · CIFAR-10 LIRA, 10 samples | ASR: 12.23 -> 5.18 | Improved |
| add MixUp augmentation to NFT (ablation) | NFT · CIFAR-10 LIRA, 10 samples | ACC: 81.05 -> 82.72 | Hurt |
| add mask scheduling function mu(l) (ablation) | NFT · CIFAR-10 LIRA, 100 samples | ASR: 4.33 -> 2.04 | Improved |
| fine-tune neural masks with MixUp instead of adversarial trigger search baselines | NFT vs AWM · CIFAR-10 LIRA, 10 samples | ASR: 91.45 -> 5.18 | Improved |
| use NFT instead of ANP/AWM adversarial search for defense | backdoor purification · CIFAR-10 runtime | runtime (sec): 118.1 -> 28.4 | Improved |
| apply NFT on ImageNet scale | NFT · ImageNet | average ASR drop / ACC drop: 94.03 / 2.84 | Improved |
| choose alpha 0.8, beta 0.50 for exponential mask schedule (ablation) | NFT · CIFAR-10 LIRA | ASR/ACC: ASR 1.53, ACC 90.57 | Improved |
| NFT in multi-label detection setting | NFT · VOC12 | ASR: RNP 13.8 vs NFT 14.2 | Mixed |
Closest papers by meaning
- Unveiling and Mitigating Backdoor Vulnerabilities based on Unlearning Weight Changes and Backdoor Activeness (NeurIPS 2024 · similarity 0.91)
- Uncovering, Explaining, and Mitigating the Superficial Safety of Backdoor Defense (NeurIPS 2024 · similarity 0.89)
- Backdoor Mitigation via Invertible Pruning Masks (NeurIPS 2025 · similarity 0.89)
- BAN: Detecting Backdoors Activated by Adversarial Neuron Noise (NeurIPS 2024 · similarity 0.88)
- From Toxic to Trustworthy: Using Self-Distillation and Semi-supervised Methods to Refine Neural Networks (AAAI 2024 · similarity 0.88)
- REFINE: Inversion-Free Backdoor Defense via Model Reprogramming (ICLR 2025 · similarity 0.87)
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.
