Gated Graph Sequence Neural Networks

17 Nov 2015Yujia LiDaniel TarlowMarc BrockschmidtRichard Zemel

Graph-structured data appears frequently in domains including chemistry, natural language semantics, social networks, and knowledge bases. In this work, we study feature learning techniques for graph-structured inputs... (read more)

PDF Abstract

Evaluation results from the paper


Task Dataset Model Metric name Metric value Global rank Compare
Node Classification CiteSeer (1%) GGNN Accuracy 56.0% # 9
Node Classification CiteSeer with Public Split: fixed 20 nodes per class GGNN Accuracy 64.6% # 14
Node Classification Cora (0.5%) GGNN Accuracy 48.2% # 10
Node Classification Cora (1%) GGNN Accuracy 60.5% # 9
Node Classification Cora (3%) GGNN Accuracy 73.1% # 9
Node Classification Cora with Public Split: fixed 20 nodes per class GGNN Accuracy 77.6% # 14
Graph Classification IPC-grounded GG-NN Accuracy 77.9% # 2
Graph Classification IPC-lifted GG-NN Accuracy 81.4% # 3
Node Classification PubMed (0.03%) GGNN Accuracy 55.8% # 10
Node Classification PubMed (0.05%) GGNN Accuracy 63.3% # 9
Node Classification PubMed (0.1%) GGNN Accuracy 70.4% # 9
Node Classification PubMed with Public Split: fixed 20 nodes per class GGNN Accuracy 75.8% # 13
Drug Discovery QM9 Gated Graph Sequence NN Error ratio 1.36 # 2
SQL-to-Text WikiSQL GGS-NN BLEU-4 35.53 # 2