AI Paper Atlas

Research map / Graph learning / GNN architectures

Fair Graph Distillation

Qizhang Feng, Zhimeng Jiang, Ruiquan Li, Yicheng Wang, Na Zou, Jiang Bian, Xia Hu

NeurIPS 2023 · Graph learning › GNN architectures · approach: Robustness, privacy, and fairness of GNNs

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 changedWhereMeasuredEffect
add FGD coherence-bias debiasing to graph distillationGCN on distilled graph · Pokec-zΔDP: 2.84±1.1% -> 1.75±1.1%Improved
add FGD debiasingGCN on distilled graph · Pokec-nΔDP: 4.02±0.6% -> 0.66±0.5%Improved
add FGD debiasingGCN on distilled graph · GermanΔDP: 5.38±3.3% -> 1.93±0.1%Improved
add FGD debiasingGCN on distilled graph · CreditΔDP: 12.04±5.4% -> 4.94±4.8%Improved
add FGD debiasingGCN on distilled graph · RecidivismΔDP: 2.68±1.0% -> 0.54±0.3%Improved
add FGD debiasingGCN on distilled graph · Pokec-naccuracy: 50.10±2.7% -> 54.80±1.5%Improved
add FGD debiasingGCN on distilled graph · Germanaccuracy: 72.83±0.8% -> 70.50±0.1%Hurt
add FGD debiasingGCN on distilled graph · Recidivismaccuracy: 70.89±1.8% -> 70.09±2.6%No clear effect
train on distilled graph instead of real graphGCN · Recidivismaccuracy: 94.45±0.0% -> 70.89±1.8%Hurt
coherence bias regularizationdistilled graph · Pokec-zcoherence bias: 0.009468 -> 0.002123Improved
coherence bias regularizationdistilled graph · Recidivismcoherence bias: 0.000098 -> 0.000038Improved
increase alpha regularization weight (ablation)FGD · RecidivismAUC and ΔDP: AUC 74.75 -> 65.35, ΔDP 0.84 -> 0.00Mixed

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy