AI Paper Atlas

Research map / Graph learning / GNN architectures

Towards Label Position Bias in Graph Neural Networks

Haoyu Han, Xiaorui Liu, Feng Shi, MohamadAli Torkamani, Charu C. Aggarwal, Jiliang Tang

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
Use LPSL-learned graph structure with APPNPAPPNP · Cora, 5 labels/classaccuracy (%): 75.86 ± 2.34 -> 77.24 ± 2.18Improved
Use LPSL-learned graph structure with APPNPAPPNP · CiteSeer, 20 labels/classaccuracy (%): 70.85 ± 1.45 -> 71.25 ± 1.14Improved
Use LPSL-learned graph structure with APPNPAPPNP · PubMed, 60% labels/classaccuracy (%): 78.93 ± 2.11 -> 87.96 ± 0.57Hurt
Use LPSL-learned graph structure with GCNGCN · Amazon Computers, 20 labels/classaccuracy (%): 79.17 ± 1.92 -> 80.05 ± 2.92Improved
Use LPSL-learned graph structure with GCNGCN · ogbn-arxiv, 54% labelsaccuracy (%): 71.91 ± 0.15 -> 72.04 ± 0.12Improved
Use LPSL-learned graph structureGCN and APPNP · Cora and CiteSeerLPS-group bias (WDP): Cora GCN 0.0494 -> 0.0361; CiteSeer APPNP 0.0344 -> 0.0275Mixed
Use LPSL-learned graph structureGCN and APPNP · Cora and CiteSeerdegree bias (WDP): Cora APPNP 0.0408 -> 0.0349; CiteSeer APPNP 0.0499 -> 0.0316Improved
Use LPSL-learned graph structureGCN and APPNP · Cora and CiteSeershortest-path-distance bias (WDP): Cora GCN 0.0237 -> 0.0150; CiteSeer GCN 0.0296 -> 0.0246Mixed
Vary smoothing parameter λ within the selected ranges (ablation)LPSL-APPNP and LPSL-GCN · Cora and CiteSeer, 20 labels/classaccuracyNo clear effect
Vary uniform proximity constraint c within the selected ranges (ablation)LPSL-APPNP and LPSL-GCN · Cora and CiteSeer, 20 labels/classaccuracyNo clear effect

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