Research map / Adversarial robustness and security / Black-box and transfer attacks
Sibling-Attack: Rethinking Transferable Adversarial Attacks Against Face Recognition
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 |
|---|---|---|---|
| Use Attribute Recognition as a sibling task for transferable impersonation attacks | Sibling-Attack · CelebA-HQ and LFW, black-box FR models | ASR: boosting ASR by 12.61% and 55.77% on average | Helped |
| Use FR and AR loss averaging as the basic multi-task framework (ablation) | Sibling-Attack · LFW, IR50 | ASR: 80.90 | Helped |
| Add hard parameter sharing (ablation) | Sibling-Attack basic framework · LFW, online FR models | ASR: 7.60% and 8.20% | Helped |
| Add joint-task meta optimization (ablation) | Sibling-Attack with hard parameter sharing · LFW, online FR models | ASR: 18.10% and 5.80% | Helped |
| Add cross-task gradient stabilization (ablation) | Sibling-Attack with JTMO · LFW, online FR models | ASR: 0.60% and 8.10% | Helped |
| Select Attribute Recognition rather than Landmark Detection or Face Parsing as the sibling task | Hard-parameter-sharing multi-task attack · CelebA-HQ and LFW, IR50 and ResNet101 | ASR: 93.00 vs. 75.20; 93.40 vs. 85.10; 97.60 vs. 75.80; 96.80 vs. 83.40 | Helped |
Closest papers by meaning
- Improving the Transferability of Adversarial Attacks on Face Recognition with Diverse Parameters Augmentation (CVPR 2025 · similarity 0.85)
- Transferable Adversarial Face Attack with Text Controlled Attribute (AAAI 2025 · similarity 0.85)
- Non-Adaptive Adversarial Face Generation (NeurIPS 2025 · similarity 0.84)
- FaceShield: Explainable Face Anti-Spoofing with Multimodal Large Language Models (AAAI 2026 · similarity 0.84)
- PubDef: Defending Against Transfer Attacks From Public Models (ICLR 2024 · similarity 0.84)
- From Intuition to Investigation: A Tool-Augmented Reasoning MLLM Framework for Generalizable Face Anti-Spoofing (CVPR 2026 · similarity 0.84)
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.