Research map / Adversarial robustness and security / Attacks on text and graphs
WARP: Weight Teleportation for Attack-Resilient Unlearning Protocols
Mohammad mahdi Maheri, Xavier Cadet, Peter Chin, Hamed Haddadi
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 WARP teleportation | NGP · U-LiRA black-box, all forget samples | TPR@1: 0.030 -> 0.014 | Improved |
| add WARP teleportation | SCRUB · U-LiRA black-box, most-memorized 1% | AUC: 0.710 -> 0.610 | Improved |
| add WARP teleportation | SF · U-LiRA black-box, most-memorized 1% | AUC: 0.518 -> 0.501 | Improved |
| add WARP teleportation | PGU · U-LiRA black-box, all forget samples | TPR@5: 0.098 -> 0.104 | Hurt |
| add WARP teleportation | BadTeacher · U-LiRA black-box, all forget samples | AUC: 0.725 -> 0.661 | Improved |
| add WARP teleportation | Salun · U-LiRA black-box, all forget samples | AUC: 0.572 -> 0.565 | Improved |
| add WARP teleportation | PGU · U-LiRA black-box, most-memorized 1% | AUC: 0.910 -> 0.875 | Improved |
| add WARP teleportation | NGP · Test accuracy | accuracy: 0.808 -> 0.797 | Hurt |
| add WARP teleportation | SCRUB · Test accuracy | accuracy: 0.815 -> 0.813 | No clear effect |
| add WARP teleportation | BadTeacher · Test accuracy | accuracy: 0.816 -> 0.818 | Improved |
| add WARP teleportation | NGP · Reconstruction attack, ImageNet-1K ResNet-18 | PSNR (dB): 10.74 -> 7.38 | Improved |
| add WARP teleportation | NGP · White-box MIA (ViT, Tiny-ImageNet) | AUC: 0.792 -> 0.755 | Improved |
Closest papers by meaning
- From Dormant to Deleted: Tamper-Resistant Unlearning Through Weight-Space Regularization (NeurIPS 2025 · similarity 0.89)
- The Utility and Complexity of In- and Out-of-Distribution Machine Unlearning (ICLR 2025 · similarity 0.88)
- Reconstruction Attacks on Machine Unlearning: Simple Models are Vulnerable (NeurIPS 2024 · similarity 0.88)
- Leveraging Per-Instance Privacy for Machine Unlearning (ICML 2025 · similarity 0.88)
- Certified Unlearning for Neural Networks (ICML 2025 · similarity 0.87)
- The Unseen Threat: Residual Knowledge in Machine Unlearning under Perturbed Samples (NeurIPS 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.
