Ru-2021-08-04-all-icsd

../../_images/pareto252.png

The current structure dataset comprises 10837 structures generated from unique ICSD prototype structures composed of single elements with zero oxidation state. A more detailed procedure is found in Phys. Rev. B 99, 214108 (2019). The procedure to estimate interatomic potentials from the dataset is found in Phys. Rev. B 99, 214108 (2019) and Phys. Rev. B 102, 174104 (2020).

Improvement from **-dataset-10000-all-icsd

  • More robust for structures with a small interatomic distance

  • More robust for structures with a large interatomic distance

  • More complex potential models are included.

  • MLPs are estimated without using DFT stress tensors.

  • MLPs are estimated by using small regression weights for energetically unstable structures.

Predictions using Pareto optimal MLPs

../../_images/prediction-ecoh-volume146.png

The cohesive energy and volume are obtained by performing a local structure optimization from the DFT equilibrium structure. In addition, the DFT equilibrium structure is obtained by optimizing a prototype structure included in ICSD, and the prototype is used as the structure legend in the figure. Therefore, the structure type of the converged structure is sometimes different from that shown in the legend even if the potential energy surface predicted by MLP is almost the same as the true one.

The other properties predicted by each Pareto optimal MLP are available from column Predictions in the following table.

Ru-2021-08-04-all-icsd shows large prediction errors. They should be carefully used. Such an MLP is often accurate for reasonable structures, but it is not accurate for unrealistic structures.

Pareto optimals

Name

Time [ms] (1core/36cores)

RMSE [meV/atom]/[eV/A]

Predictions

Files

pair-1

0.012 / 0.003

279.35 / 1.0401

–

–

pair-14

0.017 / 0.038

90.665 / 0.4678

–

–

pair-27

0.018 / 0.005

87.580 / 0.4216

–

–

pair-18

0.033 / 0.017

74.843 / 0.4623

–

–

pair-31

0.050 / 0.007

61.260 / 0.4010

–

–

pair-37

0.136 / 0.018

54.583 / 0.3829

–

–

gtinv-300

0.228 / 0.023

29.770 / 0.2344

–

–

gtinv-235

0.281 / 0.023

24.697 / 0.2118

–

–

gtinv-312

0.472 / 0.036

20.607 / 0.2128

–

–

gtinv-190

0.523 / 0.038

18.325 / 0.1981

–

–

gtinv-255

0.573 / 0.040

17.949 / 0.1964

–

–

gtinv-301

0.743 / 0.062

15.462 / 0.1784

–

–

gtinv-236

0.804 / 0.059

14.899 / 0.1698

–

–

gtinv-107

1.018 / 0.072

14.892 / 0.1816

–

–

gtinv-172

1.152 / 0.075

10.382 / 0.1487

–

–

gtinv-237

1.177 / 0.073

10.337 / 0.1482

–

–

gtinv-191

1.600 / 0.102

10.307 / 0.1558

–

–

gtinv-174

1.973 / 0.134

7.6425 / 0.1368

predictions

mlp.lammps input log

gtinv-239

2.057 / 0.133

7.6115 / 0.1365

predictions

mlp.lammps input log

gtinv-192

2.106 / 0.124

7.5226 / 0.1375

predictions

mlp.lammps input log

gtinv-257

2.164 / 0.129

7.4931 / 0.1372

predictions

mlp.lammps input log

gtinv-193

3.465 / 0.217

6.8816 / 0.1428

predictions

mlp.lammps input log

gtinv-258

3.556 / 0.220

6.8683 / 0.1425

predictions

mlp.lammps input log

gtinv-194

3.908 / 0.240

5.4997 / 0.1283

predictions

mlp.lammps input log

gtinv-288

13.824 / 0.659

5.4981 / 0.1394

predictions

mlp.lammps input log

gtinv-224

14.066 / 0.687

4.3557 / 0.1246

predictions

mlp.lammps input log

gtinv-234

21.888 / 1.017

4.3470 / 0.1271

predictions

mlp.lammps input log

Column “Time” shows the time required to compute the energy and forces for 1 MD step and 1 atom, which is estimated from a simulation of 10 runs for a structure with 284 atoms using a workstation with Intel(R) Xeon(R) CPU E5-2695 v4 @ 2.10GHz. Note that the MLPs should be carefully used for extreme structures. The MLPs often return meaningless values for them.

  • All Pareto optimal MLPs are available here.