Research map / Federated learning and privacy / Privacy attacks and secure learning
Mitigating Privacy Risk via Forget Set-Free Unlearning
Aviraj Newatia, Michael Cooper, Viet Nguyen, Rahul G Krishnan
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 |
|---|---|---|---|
| use Reload instead of GA | unlearning methods · CIFAR-100 forget 10% random | delta forget accuracy: 18.77 -> 0.30 | Improved |
| Reload (without forget set) vs forget-set methods | Reload · CIFAR-100 forget 10% random | delta forget MIA: 0.21 -> 0.01 | Improved |
| add selective re-initialisation after one ascent step (ablation) | Reload · SVHN class 8 forgetting | predicted class | Improved |
| Reload on TOFU 1% forgetting | Llama-2-7B / Phi-1.5 · TOFU 1% | model utility: +0.0748 | Improved |
| Reload on TOFU 5% and 10% forgetting | Reload · TOFU 5%/10% | model utility | Hurt |
| finetune on subsets of retain set (ablation) | Reload · unlearning | performance | Hurt |
| quantize cached gradients (ablation) | Reload · unlearning | performance | No clear effect |
Closest papers by meaning
- The Utility and Complexity of In- and Out-of-Distribution Machine Unlearning (ICLR 2025 · similarity 0.92)
- Towards Source-Free Machine Unlearning (CVPR 2025 · similarity 0.91)
- Unlearned but Not Forgotten: Data Extraction after Exact Unlearning in LLM (NeurIPS 2025 · similarity 0.91)
- Leveraging Per-Instance Privacy for Machine Unlearning (ICML 2025 · similarity 0.90)
- Certified Unlearning for Neural Networks (ICML 2025 · similarity 0.90)
- Learning to Rewind via Iterative Prediction of Past Weights for Practical Unlearning (AAAI 2025 · similarity 0.90)
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.
