AI Paper Atlas

Research map / Graph learning / GNN architectures

Dirichlet Energy Constrained Learning for Deep Graph Neural Networks

Kaixiong Zhou, Xiao Huang, Daochen Zha, Rui Chen, Li Li, Soo-Hyun Choi, Xia Hu

NeurIPS 2021 · Graph learning › GNN architectures · approach: Simple and scalable GNNs · 38 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
Use EGNN instead of the strongest reported baseline at 64 layersCora node classification · Cora, 64 layersaccuracy (%): 85.4 -> 85.7Improved
Use EGNN instead of the strongest reported baseline at 64 layersPubmed node classification · Pubmed, 64 layersaccuracy (%): 79.9 -> 80.1Improved
Use EGNN instead of the strongest reported baseline at 32 layersCoauthor-Physics node classification · Coauthor-Physics, 32 layersaccuracy (%): 92.9 -> 93.3Improved
Use EGNN instead of the strongest reported baseline at 32 layersOgbn-arxiv node classification · Ogbn-arxiv, 32 layersaccuracy (%): 71.4 -> 72.7Improved
Replace orthogonal weight initialization and regularization with Glorot initialization and Frobenius norm regularization (ablation)EGNN · Cora, 64 layersaccuracy (%): 85.7 -> 23.6Hurt
Replace orthogonal weight initialization and regularization with Glorot initialization and Frobenius norm regularization (ablation)EGNN · Ogbn-arxiv, 32 layersaccuracy (%): 72.7 -> 72.7No clear effect
Remove the lower energy limit by setting c_min to 0 (ablation)EGNN · Cora, 64 layersaccuracy (%): 85.7 -> 12.9Hurt
Increase the lower limit c_min to 0.95 (ablation)EGNN · Cora, 64 layersaccuracy (%): 85.7 -> 71.5Hurt
Replace SReLU with linear identity activation (ablation)EGNN · Cora, 64 layersaccuracy (%): 85.7 -> 85.5No clear effect
Replace SReLU with ReLU (ablation)EGNN · Ogbn-arxiv, 32 layersaccuracy (%): 72.7 -> 72.4Hurt

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