FIG. 2. Performance evaluation of the universal materials model trained by the DeepH method using a large materials database comprising ~104 solid materials. (IMAGE)

Science China Press

FIG. 2. Performance evaluation of the universal materials model trained by the DeepH method using a large materials database comprising ~104 solid materials.

Caption

Mean absolute error (MAE)s of the predicted DFT Hamiltonian averaged for each element, showing high prediction accuracy at the level of meV.

Credit

©Science China Press

Usage Restrictions

Use with credit.

License

Original content

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.