Improving the Effective Receptive Field of Message-Passing Neural Networks

Benchmark Model Rank Results
node-classification-on-coco-spIM-GatedGCN#5macro F1: 0.3501 ± 0.0061
node-classification-on-coco-spIM-GCN#15macro F1: 0.1960 ± 0.0023
node-classification-on-log-angelesIM-GCN#1Average Top-1 Accuracy: 62.4 ± 0.3
node-classification-on-londonIM-GCN#1Average Top-1 Accuracy: 58.9 ± 0.1
node-classification-on-parisIM-GCN#1Average Top-1 Accuracy: 55.3 ± 0.3
node-classification-on-pascalvoc-sp-1IM-GatedGCN#4macro F1: 0.4332 ± 0.0078
node-classification-on-pascalvoc-sp-1IM-GCN#13macro F1: 0.2929 ± 0.0058
node-classification-on-shanghaiIM-GCN#1Average Top-1 Accuracy: 67.8 ± 0.1