Research map / Learning theory and algorithms / Fairness
Beyond Binary Erasure: Soft-Weighted Unlearning for Fairness and Robustness
Xinbao Qiao, Ningning Ding, Yushi Cheng, Meng Zhang
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 |
|---|---|---|---|
| replace hard binary removal with soft sample weights (ablation) | Gradient Ascent+FT · Bank (LR) | fairness: 98.9% improvement | Improved |
| replace hard binary removal with soft sample weights (ablation) | IF · Bank (LR) | fairness: 97.2% improvement | Improved |
| replace hard removal with soft weights (ablation) | Bad-T · Jigsaw (LR) | fairness: -0.7% | Hurt |
| replace hard removal with soft weights (ablation) | Fisher · Adult (NN) | fairness: -3.4% | Hurt |
| replace hard removal with soft weights (ablation) | EU-k · Adult (NN) | fairness: 84.4% | Improved |
| replace hard removal with soft weights (ablation) | nine unlearning algorithms · fairness tasks | fairness: up to 98.9% | Improved |
| replace hard removal with soft weights (ablation) | IF · Adult (LR) | utility: 5.6% (fairness task) | Improved |
| add soft-weight step instead of binary assignment (ablation) | IF · runtime | execution time: 0.03% of total execution time | No clear effect |
| use influence estimates versus actual leave-one-out changes | influence function · Adult | correlation: Spearman 1.00, Pearson 0.97 | Improved |
Closest papers by meaning
- The Utility and Complexity of In- and Out-of-Distribution Machine Unlearning (ICLR 2025 · similarity 0.91)
- Certified Unlearning for Neural Networks (ICML 2025 · similarity 0.90)
- Langevin Unlearning: A New Perspective of Noisy Gradient Descent for Machine Unlearning (NeurIPS 2024 · similarity 0.89)
- Rewind-to-Delete: Certified Machine Unlearning for Nonconvex Functions (NeurIPS 2025 · similarity 0.89)
- A Certified Unlearning Approach without Access to Source Data (ICML 2025 · similarity 0.89)
- What makes unlearning hard and what to do about it (NeurIPS 2024 · similarity 0.89)
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.
