Research map / Adversarial robustness and security / Attacks on text and graphs
T-BFA: T argeted B it- F lip Adversarial Weight A ttack
Adnan Siraj Rakin, Zhezhi He, Jingtao Li, Fan Yao, Chaitali Chakrabarti, Deliang Fan
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 |
|---|---|---|---|
| target a specific class (N-to-1 T-BFA) instead of untargeted BFA | BFA · CIFAR-10 ResNet-20 8-bit | # of bit-flips: 28 -> 4 | Improved |
| lower weight quantization from 8-bit to 2-bit (ablation) | T-BFA N-to-1 · CIFAR-10 ResNet-20 | # of bit-flips: 6.0 -> 37.6 | Hurt |
| lower weight quantization from 8-bit to 2-bit (ablation) | T-BFA 1-to-1 · CIFAR-10 ResNet-20 | # of bit-flips: 2.2 -> 17.6 | Hurt |
| lower weight quantization from 8-bit to 2-bit (ablation) | T-BFA 1-to-1 (S) · CIFAR-10 ResNet-20 | post-attack test accuracy (%): 80.3 -> 84.6 | Improved |
| defend with piece-wise clustering (PC) (ablation) | 8-bit ResNet-20 · CIFAR-10 1-to-1 (S) | # of bit-flips: 3.6 -> 8.6 | Mixed |
| defend with binarization (ablation) | ResNet-20 · CIFAR-10 1-to-1 (S) | # of bit-flips: 3.6 -> 16 | Mixed |
| defend with binarization (ablation) | ResNet-20 · CIFAR-10 N-to-1 | # of bit-flips: 5.2 -> 35.5 | Mixed |
| increase network capacity (VGG-11 instead of ResNet-20) | T-BFA N-to-1 · CIFAR-10 | # of bit-flips: 5.1 -> 3.0 | Hurt |
| use compact network MobileNet-V2 instead of ResNet-18/34 | T-BFA 1-to-1 (S) · ImageNet | post-attack test accuracy (%): 59.48 -> 33.99 | Hurt |
| protect the entire last classification layer (ablation) | T-BFA · CIFAR-10 | # of bit-flips: all less than 30 additional flips | No clear effect |
| 1-to-1 (S) instead of TBT | T-BFA · CIFAR-10 ResNet-20 8-bit | # of bit-flips: 199 -> 12.2 | Improved |
Closest papers by meaning
- Defending and Harnessing the Bit-Flip Based Adversarial Weight Attack (CVPR 2020 · similarity 0.92)
- Your Scale Factors are My Weapon: Targeted Bit-Flip Attacks on Vision Transformers via Scale Factor Manipulation (CVPR 2025 · similarity 0.85)
- Crossfire: An Elastic Defense Framework for Graph Neural Networks Under Bit Flip Attacks (AAAI 2025 · similarity 0.84)
- Constrained Adaptive Attack: Effective Adversarial Attack Against Deep Neural Networks for Tabular Data (NeurIPS 2024 · similarity 0.83)
- TtBA: Two-third Bridge Approach for Decision-Based Adversarial Attack (ICML 2025 · similarity 0.82)
- Enhancing Adversarial Transferability with Adversarial Weight Tuning (AAAI 2025 · similarity 0.82)
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.
