Cd-2021-08-04-all-icsd¶

The current structure dataset comprises 12826 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-volume121.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.

Pareto optimals¶
Name	Time [ms] (1core/36cores)	RMSE [meV/atom]/[eV/A]	Predictions	Files
pair-14	0.017 / 0.038	16.323 / 0.0685	–	–
pair-27	0.018 / 0.005	14.405 / 0.0668	–	–
pair-28	0.027 / 0.007	12.909 / 0.0527	–	–
pair-29	0.039 / 0.009	10.715 / 0.0503	–	–
pair-32	0.051 / 0.010	9.7951 / 0.0486	predictions	`mlp.lammps` `input` `log`
pair-33	0.069 / 0.012	9.5609 / 0.0478	predictions	`mlp.lammps` `input` `log`
pair-34	0.096 / 0.016	8.9987 / 0.0472	predictions	`mlp.lammps` `input` `log`
pair-37	0.136 / 0.018	8.3333 / 0.0458	predictions	`mlp.lammps` `input` `log`
pair-38	0.171 / 0.022	8.0639 / 0.0449	predictions	`mlp.lammps` `input` `log`
gtinv-300	0.228 / 0.023	3.3440 / 0.0212	predictions	`mlp.lammps` `input` `log`
gtinv-235	0.281 / 0.023	2.9656 / 0.0193	predictions	`mlp.lammps` `input` `log`
gtinv-312	0.472 / 0.036	2.3859 / 0.0163	predictions	`mlp.lammps` `input` `log`
gtinv-190	0.523 / 0.038	2.1580 / 0.0151	predictions	`mlp.lammps` `input` `log`
gtinv-255	0.573 / 0.040	2.1055 / 0.0149	predictions	`mlp.lammps` `input` `log`
gtinv-195	0.748 / 0.053	2.0771 / 0.0153	predictions	`mlp.lammps` `input` `log`
gtinv-260	0.787 / 0.053	2.0299 / 0.0149	predictions	`mlp.lammps` `input` `log`
gtinv-321	1.240 / 0.088	2.0281 / 0.0162	predictions	`mlp.lammps` `input` `log`
gtinv-313	1.590 / 0.102	1.8869 / 0.0143	predictions	`mlp.lammps` `input` `log`
gtinv-191	1.600 / 0.102	1.7978 / 0.0134	predictions	`mlp.lammps` `input` `log`
gtinv-256	1.780 / 0.109	1.7734 / 0.0132	predictions	`mlp.lammps` `input` `log`
gtinv-220	1.890 / 0.117	1.6563 / 0.0127	predictions	`mlp.lammps` `input` `log`
gtinv-346	3.371 / 0.155	1.4568 / 0.0141	predictions	`mlp.lammps` `input` `log`
gtinv-351	3.420 / 0.142	1.3826 / 0.0125	predictions	`mlp.lammps` `input` `log`
gtinv-221	5.913 / 0.339	1.3818 / 0.0112	predictions	`mlp.lammps` `input` `log`
gtinv-286	6.000 / 0.346	1.3721 / 0.0111	predictions	`mlp.lammps` `input` `log`
gtinv-222	6.837 / 0.369	1.3163 / 0.0109	predictions	`mlp.lammps` `input` `log`
gtinv-227	8.349 / 0.454	1.3130 / 0.0110	predictions	`mlp.lammps` `input` `log`
gtinv-332	13.113 / 0.653	1.3079 / 0.0113	predictions	`mlp.lammps` `input` `log`
gtinv-223	13.159 / 0.653	1.2770 / 0.0107	predictions	`mlp.lammps` `input` `log`
gtinv-288	13.824 / 0.659	1.2696 / 0.0106	predictions	`mlp.lammps` `input` `log`
gtinv-224	14.066 / 0.687	1.2296 / 0.0104	predictions	`mlp.lammps` `input` `log`
gtinv-293	16.676 / 0.807	1.2278 / 0.0106	predictions	`mlp.lammps` `input` `log`
gtinv-229	17.857 / 0.829	1.2059 / 0.0104	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.