AI Paper Atlas

Research map / Graph learning / GNN architectures

Optimization and Generalization Analysis of Transduction through Gradient Boosting and Application to Multi-scale Graph Neural Networks

Kenta Oono, Taiji Suzuki

NeurIPS 2020 · Graph learning › GNN architectures · approach: GNN generalization theory · 9 citations (OpenAlex)

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 end-to-end fine-tuning to GB-GNN-Adj (ablation)GB-GNN-Adj · Coraaccuracy (%): 79.9 \pm 0.8 -> 80.4 \pm 0.8Improved
Add end-to-end fine-tuning to GB-GNN-Adj (ablation)GB-GNN-Adj · CiteSeeraccuracy (%): 70.5 \pm 0.8 -> 70.8 \pm 0.8Improved
Add end-to-end fine-tuning to GB-GNN-Adj (ablation)GB-GNN-Adj · PubMedaccuracy (%): 79.4 \pm 0.2 -> 79.0 \pm 0.5Hurt
Add end-to-end fine-tuning to GB-GNN-KTA (ablation)GB-GNN-KTA · Coraaccuracy (%): 80.9 \pm 0.9 -> 82.3 \pm 1.1Improved
Add end-to-end fine-tuning to GB-GNN-KTA (ablation)GB-GNN-KTA · CiteSeeraccuracy (%): 73.1 \pm 1.1 -> 70.8 \pm 1.0Hurt
Add end-to-end fine-tuning to GB-GNN-KTA (ablation)GB-GNN-KTA · PubMedaccuracy (%): 79.1 \pm 0.4 -> N.A.Mixed
Use KTA aggregation instead of augmented-adjacency aggregationGB-GNN · Coraaccuracy (%): 79.9 \pm 0.8 -> 80.9 \pm 0.9Improved
Use KTA aggregation instead of augmented-adjacency aggregationGB-GNN · CiteSeeraccuracy (%): 70.5 \pm 0.8 -> 73.1 \pm 1.1Improved
Use KTA aggregation instead of augmented-adjacency aggregationGB-GNN · PubMedaccuracy (%): 79.4 \pm 0.2 -> 79.1 \pm 0.4Hurt

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