Voronoi RF

Version: v1.1.2 Added: 2022-11-26 Published: 2017-07-14 26.2M parameters Missing preds: 2 (0.001%) pip install scikit-learn
Repo Paper DOI Files PyPI PR Checkpoint
Predictions
Discovery

ML vs DFT Formation Energies

ML vs DFT Convex Hull Distance

Convex hull distance prediction errors projected onto elements

1 H 0.59
2 He 0.00
3 Li 0.24
4 Be 0.37
5 B 0.49
6 C 0.36
7 N 0.69
8 O 0.87
9 F 1.00
10 Ne 0.00
11 Na 0.34
12 Mg 0.27
13 Al 0.30
14 Si 0.33
15 P 0.35
16 S 0.49
17 Cl 0.68
18 Ar 0.00
19 K 0.30
20 Ca 0.32
21 Sc 0.27
22 Ti 0.29
23 V 0.40
24 Cr 0.52
25 Mn 0.41
26 Fe 0.46
27 Co 0.33
28 Ni 0.26
29 Cu 0.30
30 Zn 0.30
31 Ga 0.27
32 Ge 0.31
33 As 0.34
34 Se 0.49
35 Br 0.66
36 Kr 0.00
37 Rb 0.29
38 Sr 0.32
39 Y 0.26
40 Zr 0.28
41 Nb 0.33
42 Mo 0.30
43 Tc 0.25
44 Ru 0.44
45 Rh 0.28
46 Pd 0.27
47 Ag 0.28
48 Cd 0.29
49 In 0.36
50 Sn 0.29
51 Sb 0.30
52 Te 0.46
53 I 0.54
54 Xe 0.08
55 Cs 0.28
56 Ba 0.30
57 La 0.26
58 Ce 0.27
59 Pr 0.27
60 Nd 0.26
61 Pm 0.22
62 Sm 0.26
63 Eu 0.29
64 Gd 0.25
65 Tb 0.28
66 Dy 0.28
67 Ho 0.28
68 Er 0.29
69 Tm 0.30
70 Yb 0.35
71 Lu 0.25
72 Hf 0.33
73 Ta 0.41
74 W 0.31
75 Re 0.26
76 Os 0.45
77 Ir 0.34
78 Pt 0.28
79 Au 0.34
80 Hg 0.26
81 Tl 0.36
82 Pb 0.40
83 Bi 0.29
84 Po 0.00
85 At 0.00
86 Rn 0.00
87 Fr 0.00
88 Ra 0.00
89 Ac 0.23
90 Th 0.31
91 Pa 0.30
92 U 0.46
93 Np 0.50
94 Pu 0.39
95 Am 0.00
96 Cm 0.00
97 Bk 0.00
98 Cf 0.00
99 Es 0.00
100 Fm 0.00
101 Md 0.00
102 No 0.00
103 Lr 0.00
104 Rf 0.00
105 Db 0.00
106 Sg 0.00
107 Bh 0.00
108 Hs 0.00
109 Mt 0.00
110 Ds 0.00
111 Rg 0.00
112 Cn 0.00
113 Nh 0.00
114 Fl 0.00
115 Mc 0.00
116 Lv 0.00
117 Ts 0.00
118 Og 0.00
57-71 La-Lu Lanthanides
89-103 Ac-Lr Actinides

Model Authors

  1. Logan Ward  Argonne National Laboratory  
  2. Chris Wolverton  Northwestern University  

Trained By

  1. Janosh Riebesell  University of Cambridge, Lawrence Berkeley National Laboratory  

Model Info

  • Model Version v1.1.2
  • Model Type Fingerprint
  • Targets E
  • Openness OSOD
  • Train Task RS2RE
  • Test Task IS2E
  • Trained for Benchmark Yes

Training Set

MP v2022.10.28: 155k structures

Description

A random forest trained to map the combo of composition-based Magpie features and structure-based relaxation-invariant Voronoi tessellation features (bond angles, coordination numbers, ...) to DFT formation energies.

Long

This is an old model that predates most deep learning for materials but significantly improved over Coulomb matrix and partial radial distribution function methods. It therefore serves as a good baseline model to see what modern ML buys us.

Dependencies