AI Paper Atlas

Research map / Graph learning / GNN architectures

Boost then Convolve: Gradient Boosting Meets Graph Neural Networks

Sergei Ivanov, Liudmila Prokhorenkova

ICLR 2021 · Graph learning › GNN architectures · approach: Simple and scalable GNNs · 27 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
Train GBDT and GNN end-to-end, fitting new trees to GNN feature-gradient updatesGAT-based BGNN · House, County, VK, AvazuRMSE: 0.54 -> 0.5; 1.45 -> 1.26; 7.22 -> 6.95; 0.1134 -> 0.109Improved
Train GBDT and GNN end-to-end, fitting new trees to GNN feature-gradient updatesGAT-based BGNN · House_class, VK_classaccuracy: 0.625 -> 0.682; 0.577 -> 0.683Improved
Train GBDT and GNN jointlyGAT-based BGNN · SLAP, DBLPaccuracy: 0.963 -> 0.95; 0.913 -> 0.889Mixed
Train GBDT and GNN jointlyGAT-based BGNN · WikiRMSE: 45916 -> 49222Hurt
Train GBDT and GNN jointlyGAT-based BGNN · OGB-ArXivaccuracy: 0.70 -> 0.67Hurt
Use end-to-end BGNN rather than pretrained-GBDT Res-GNNGAT-based GBDT-GNN · House, County, VK, AvazuRMSE: 0.51 -> 0.5; 1.33 -> 1.26; 7.07 -> 6.95; 0.1095 -> 0.109Improved
Use end-to-end BGNN rather than pretrained-GBDT Res-GNNGAT-based GBDT-GNN · House_class, VK_classaccuracy: 0.625 -> 0.682; 0.603 -> 0.683Improved
Use end-to-end BGNN rather than pretrained-GBDT Res-GNNGAT-based GBDT-GNN · SLAP, DBLPaccuracy: 0.905 -> 0.95; 0.892 -> 0.889Mixed
Use end-to-end BGNN rather than pretrained-GBDT Res-GNNGAT-based GBDT-GNN · OGB-ArXivaccuracy: 0.70 -> 0.67Hurt
Add GBDT preprocessing to GATRes-GNN · House regressionRMSE: 0.54 -> 0.51Improved
Add GBDT preprocessing to GATRes-GNN · Wiki regressionRMSE: 45916 -> 46747Hurt
Use BGNN instead of GATGAT · House regressiontraining time: 35 $\pm$ 2 -> 20 $\pm$ 4Improved

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