Research map / Graph learning / Graph self-supervised learning
GUIC: Certified Graph Unlearning with Individual Fairness Guarantees
Zichong Wang, Tongliang Liu, Wenbin 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 |
|---|---|---|---|
| add individual-fairness-aware proxy node and orthogonal fairness-restoration update to certified unlearning | Certified graph unlearning (SGC) · Bail | FNDCG: 57.37 -> 71.69 (vs Certified) | Improved |
| add individual-fairness-aware proxy node and orthogonal fairness-restoration update | Certified graph unlearning (SGC) · Pokec-z | FNDCG: 49.04 -> 65.81 (vs Certified) | Improved |
| add individual-fairness-aware proxy node and orthogonal fairness-restoration update | Certified graph unlearning (SGC) · Pokec-n | FNDCG: 42.94 -> 61.06 (vs Certified) | Improved |
| add fairness-preserving update to unlearning | Certified unlearning · Bail, Pokec-z, Pokec-n | Accuracy: Bail 79.80 -> 78.70; Pokec-z 68.73 -> 67.87; Pokec-n 64.37 -> 63.23 | Mixed |
| add individual fairness component (GUIC vs GUIC-NI) (ablation) | GUIC · Bail, Pokec-z, Pokec-n | FNDCG | Improved |
| add individual fairness component (GUIC vs GUIC-NI) (ablation) | GUIC · Bail, Pokec-z, Pokec-n | Accuracy and F1 | No clear effect |
| add individual fairness computations to unlearning (GUIC vs GUIC-NI) (ablation) | GUIC · Bail, Pokec-z, Pokec-n | unlearning runtime: slightly slower than GUIC-NI | Hurt |
| use GUIC instead of retraining from scratch | GUIC · Bail, Pokec-z, Pokec-n | runtime: at least 3x faster | Improved |
| use purely performance-driven or privacy-driven unlearning (L-CODEC, SGC, Certified) | graph unlearning · Pokec-z | FNDCG: 68.04 (Original) -> 43.33 (L-CODEC) | Hurt |
Closest papers by meaning
- Erase Then Rectify: A Training-Free Parameter Editing Approach for Cost-Effective Graph Unlearning (AAAI 2025 · similarity 0.90)
- Towards Effective and General Graph Unlearning via Mutual Evolution (AAAI 2024 · similarity 0.89)
- Scalable and Certifiable Graph Unlearning: Overcoming the Approximation Error Barrier (ICLR 2025 · similarity 0.88)
- Fair Graph U-Net: A Fair Graph Learning Framework Integrating Group and Individual Awareness (AAAI 2025 · similarity 0.88)
- Community-Centric Graph Unlearning (AAAI 2025 · similarity 0.87)
- DANCE: Dual Unbiased Expansion with Group-acquired Alignment for Out-of-distribution Graph Fairness Learning (ICML 2025 · similarity 0.86)
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.
