Table 1 Comparison of ML models for space group and point group one-hot classification problems

From: Faithful novel machine learning for predicting quantum properties

Model

Point Group Accuracy

Space Group Accuracy

Baseline

0.15

0.10

naive Neural Network (NNN)

0.14

0.14

CGNN

0.78

0.78

CCNN

0.81

0.79

CANN

0.73

0.62

  1. CCNN achieved the highest performance, indicating an alternative way forward for structural classification.
Back to article page