Pt-2021-08-04-all-icsd

../../_images/pareto248.png

The current structure dataset comprises 12266 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-volume142.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.

Pt-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-14

0.017 / 0.038

51.208 / 0.2643

pair-27

0.018 / 0.005

36.938 / 0.1783

pair-28

0.027 / 0.007

35.515 / 0.1772

pair-31

0.050 / 0.007

34.115 / 0.1696

pair-32

0.051 / 0.010

30.532 / 0.1675

pair-33

0.069 / 0.012

29.838 / 0.1654

pair-37

0.136 / 0.018

24.212 / 0.1575

pair-38

0.171 / 0.022

23.776 / 0.1569

gtinv-300

0.228 / 0.023

11.379 / 0.1214

gtinv-235

0.281 / 0.023

9.2120 / 0.1088

predictions

mlp.lammps input log

gtinv-175

0.376 / 0.031

9.2081 / 0.1146

predictions

mlp.lammps input log

gtinv-240

0.393 / 0.030

9.1919 / 0.1144

predictions

mlp.lammps input log

gtinv-312

0.472 / 0.036

8.3851 / 0.1131

predictions

mlp.lammps input log

gtinv-190

0.523 / 0.038

7.6308 / 0.1062

predictions

mlp.lammps input log

gtinv-255

0.573 / 0.040

7.0458 / 0.1032

predictions

mlp.lammps input log

gtinv-301

0.743 / 0.062

6.3824 / 0.1024

predictions

mlp.lammps input log

gtinv-236

0.804 / 0.059

5.8926 / 0.0978

predictions

mlp.lammps input log

gtinv-176

1.108 / 0.075

5.7824 / 0.0976

predictions

mlp.lammps input log

gtinv-241

1.126 / 0.076

5.7548 / 0.0975

predictions

mlp.lammps input log

gtinv-172

1.152 / 0.075

4.5699 / 0.0896

predictions

mlp.lammps input log

gtinv-237

1.177 / 0.073

4.5673 / 0.0895

predictions

mlp.lammps input log

gtinv-256

1.780 / 0.109

4.4107 / 0.0899

predictions

mlp.lammps input log

gtinv-174

1.973 / 0.134

3.9402 / 0.0849

predictions

mlp.lammps input log

gtinv-239

2.057 / 0.133

3.9165 / 0.0849

predictions

mlp.lammps input log

gtinv-192

2.106 / 0.124

3.8241 / 0.0859

predictions

mlp.lammps input log

gtinv-257

2.164 / 0.129

3.7219 / 0.0842

predictions

mlp.lammps input log

gtinv-193

3.465 / 0.217

3.6587 / 0.0856

predictions

mlp.lammps input log

gtinv-258

3.556 / 0.220

3.4198 / 0.0837

predictions

mlp.lammps input log

gtinv-194

3.908 / 0.240

3.1728 / 0.0795

predictions

mlp.lammps input log

gtinv-259

4.078 / 0.244

3.1518 / 0.0795

predictions

mlp.lammps input log

gtinv-264

5.489 / 0.315

3.0867 / 0.0800

predictions

mlp.lammps input log

gtinv-222

6.837 / 0.369

3.0429 / 0.0807

predictions

mlp.lammps input log

gtinv-223

13.159 / 0.653

2.8467 / 0.0813

predictions

mlp.lammps input log

gtinv-224

14.066 / 0.687

2.5423 / 0.0758

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.