Geometric deep learning on graphs and manifolds using mixture model CNNs

Benchmark Model Rank Results
superpixel-image-classification-on-75-superpixel-mnistMonet#5Classification Error: 8.89
node-classification-on-pattern-100kMoNet#6Accuracy (%): 85.482
graph-regression-on-zinc-100kMoNet#7MAE: 0.407
graph-classification-on-cifar10-100kMoNet#19Accuracy (%): 53.42
graph-regression-on-zinc-500kMoNet#32MAE: 0.292