CoChat
Sign Up
Cited by 884Machine Learning in Materials ScienceAdvanced Chemical Physics StudiesNeural Networks and Applications

The DOI matches its Crossref record; 6 of the 6 references it cites were found in a registry or an index.

Constructing high‐dimensional neural network potentials: A tutorial review

Open access
Behler, Jörg·2015·International Journal of Quantum Chemistry
DOI 10.1002/qua.24890
Open this paper in CoChat — free accountOpen PDF

Abstract

A lot of progress has been made in recent years in the development of atomistic potentials using machine learning (ML) techniques. In contrast to most conventional potentials, which are based on physical approximations and simplifications to derive an analytic functional relation between the atomic configuration and the potential‐energy, ML potentials rely on simple but very flexible mathematical terms without a direct physical meaning. Instead, in case of ML potentials the topology of the potential‐energy surface is “learned” by adjusting a number of parameters with the aim to reproduce a set of reference electronic structure data as accurately as possible. Due to this bias‐free construction, they are applicable to a wide range of systems without changes in their functional form, and a very high accuracy close to the underlying first‐principles data can be obtained. Neural network potentials (NNPs), which have first been proposed about two decades ago, are an important class of ML potentials. Although the first NNPs have been restricted to small molecules with only a few degrees of freedom, they are now applicable to high‐dimensional systems containing thousands of atoms, which enables addressing a variety of problems in chemistry, physics, and materials science. In this tutorial review, the basic ideas of NNPs are presented with a special focus on developing NNPs for high‐dimensional condensed systems. A recipe for the construction of these potentials is given and remaining limitations of the method are discussed. © 2015 Wiley Periodicals, Inc.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1002/qua.24890

Cites

6 of 6 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Using an ensemble of models is a simple, popular choice for providing predictive uncertainty and improving generalization of machine-learning predictions.

    “Using an ensemble of models is a simple, popular choice for providing predictive uncertainty and improving generalization of machine-learning predictions33.”

    From Scaling deep learning for materials discovery (2023) · cites this paper as [33]

    In “Selecting the training data” (page unavailable), the review describes comparing predictions from several neural-network potentials to identify structures where predictions disagree, then adding training data to improve the potential. This supports using multiple models to flag unreliable predictions, but it does not say that ensembles themselves improve generalization or that they are a popular choice for predictive uncertainty. (LLM, conf=0.85)

    Experimental — accuracy not yet measured

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Flagged first, then by citations

Research works printed but not linked: 106

  1. 10.1002/andp.19273892002 · DOI
  2. 10.1103/PhysRevLett.55.2471 · DOI
  3. 10.1017/CBO9780511609633 · DOI
  4. Parr R. G., 1989, Density Functional Theory of Atoms and Molecules
  5. 10.1002/3527600043 · DOI
  6. 10.1021/ja00205a001 · DOI
  7. 10.1021/ja00124a002 · DOI
  8. 10.1021/j100389a010 · DOI
  9. 10.1021/ja00051a040 · DOI
  10. 10.1002/jcc.540040211 · DOI
  11. 10.1103/PhysRevB.31.5262 · DOI
  12. 10.1103/PhysRevLett.56.632 · DOI
  13. 10.1016/j.pmatsci.2006.10.003 · DOI
  14. 10.1021/jp004368u · DOI
  15. 10.1098/rspa.1939.0006 · DOI
  16. 10.1063/1.1622379 · DOI
  17. 10.1039/c0cp02722g · DOI
  18. 10.1063/1.466801 · DOI
  19. 10.1016/S0009-2614(03)01033-9 · DOI
  20. 10.1103/PhysRevLett.104.136403 · DOI · Paper page
  21. 10.1103/PhysRevB.88.054104 · DOI
  22. 10.1021/ct400195d · DOI
  23. 10.1007/s00214-012-1137-7 · DOI
  24. 10.1016/S0009-2614(99)00881-7 · DOI
  25. 10.1063/1.2730798 · DOI
  26. 10.1063/1.469597 · DOI
  27. 10.1016/j.cplett.2004.07.076 · DOI
  28. A.Vitek M.Stacho P.Kromer V.Snasel In: 5th IEEE International Conference on Intelligent Networking and Collaborative Systems (INCoS) 2013 pp.121–126. · DOI
  29. 10.1140/epjb/e2014-50070-0 · DOI
  30. Bishop C. M., 1996, Neural Networks for Pattern Recognition
  31. Haykin S., 2009, Neural Networks and Learning Machines
  32. 10.1002/anie.199305031 · DOI
  33. 10.1016/0893-6080(88)90020-2 · DOI
  34. 10.1007/BF02551274 · DOI · Paper page
  35. 10.1016/S0893-6080(97)00097-X · DOI
  36. 10.1016/S0893-6080(97)00010-5 · DOI
  37. 10.1016/0893-6080(91)90009-T · DOI · Paper page
  38. 10.1088/0953-8984/26/18/183001 · DOI
  39. 10.1016/S0009-2614(98)00207-3 · DOI
  40. 10.1063/1.477550 · DOI
  41. 10.1016/S0009-2614(97)00448-X · DOI
  42. 10.1063/1.3159748 · DOI
  43. 10.1063/1.2918503 · DOI
  44. 10.1063/1.2794338 · DOI
  45. 10.1063/1.2746232 · DOI
  46. 10.1103/PhysRevB.73.115431 · DOI
  47. 10.1103/PhysRevLett.94.036104 · DOI
  48. 10.1103/PhysRevB.77.115421 · DOI
  49. 10.1103/PhysRevLett.101.096104 · DOI
  50. 10.1103/PhysRevB.81.035410 · DOI
  51. 10.1088/1367-2630/14/1/013050 · DOI
  52. 10.1103/PhysRevLett.112.156101 · DOI
  53. 10.1016/j.jelechem.2008.07.032 · DOI
  54. 10.1016/j.susc.2009.12.025 · DOI
  55. 10.1007/s11426-013-5005-7 · DOI
  56. 10.1016/S0022-2860(02)00299-5 · DOI
  57. 10.1021/jp972209d · DOI
  58. 10.1021/jp9105585 · DOI
  59. 10.1088/0965-0393/7/3/308 · DOI
  60. 10.1016/S0168-583X(99)00057-9 · DOI
  61. 10.1016/j.nimb.2006.11.040 · DOI
  62. 10.1088/0953-8984/20/28/285219 · DOI
  63. 10.1063/1.2336223 · DOI
  64. 10.1063/1.2387950 · DOI
  65. 10.1063/1.2746846 · DOI
  66. 10.1063/1.3021471 · DOI
  67. 10.1016/j.cpc.2009.05.022 · DOI
  68. 10.1103/PhysRevLett.98.146401 · DOI · Paper page
  69. 10.1063/1.3553717 · DOI
  70. 10.1103/PhysRevLett.100.185501 · DOI
  71. 10.1002/pssb.200844219 · DOI
  72. 10.1103/PhysRevB.85.045439 · DOI
  73. 10.1103/PhysRevB.81.100103 · DOI
  74. 10.1038/nmat3078 · DOI
  75. 10.1103/PhysRevB.81.184107 · DOI
  76. 10.1103/PhysRevLett.108.115701 · DOI
  77. 10.1103/PhysRevB.83.153101 · DOI
  78. 10.1002/pssb.201248370 · DOI
  79. 10.1063/1.3682557 · DOI
  80. 10.1021/jp401225b · DOI
  81. 10.1524/zpch.2013.0384 · DOI
  82. 10.1039/C4CP04751F. · DOI
  83. 10.1103/PhysRevB.85.174103 · DOI
  84. 10.1103/PhysRevB.86.104301 · DOI
  85. 10.1002/pssb.201200355 · DOI
  86. 10.1021/jz402268v · DOI
  87. 10.1021/ct049976i · DOI
  88. 10.1063/1.3095491 · DOI
  89. 10.1016/S0097-8485(98)00008-4 · DOI
  90. 10.1002/andp.19213690304 · DOI
  91. 10.1063/1.2741258 · DOI
  92. 10.1039/b608822h · DOI
  93. 10.1007/BF00549096 · DOI
  94. 10.1002/qua.21507 · DOI
  95. 10.1021/ct800166r · DOI
  96. 10.1002/wcms.30 · DOI · Paper page
  97. 10.1063/1.3382344 · DOI · Paper page
  98. 10.1080/01442350903234923 · DOI
  99. 10.1063/1.4825111 · DOI
  100. 10.1063/1.4712397 · DOI
  101. 10.1109/37.55119 · DOI
  102. 10.1073/pnas.202427399 · DOI · Paper page
  103. Press W. H., 2007, Numerical Recipes ‐ The Art of Scientific Computing
  104. 10.1115/1.3662552 · DOI · Paper page
  105. 10.1002/0471221546 · DOI
  106. 10.1002/cem.1180080605 · DOI

Cited by

Sorted by citations

Metadata

Authors
Behler, J.
Affiliations
Ruhr University Bochum (DE)
Corresponding author
Behler, Jörg
ORCID
Behler, Jörg
Venue
International Journal of Quantum Chemistry
Publisher
Wiley
Published
2015-03-06
Volume
115
Issue
16
Pages
1032-1050
ISSN
  • 0020-7608 (ISSN-L)
  • 1097-461X
DOI
10.1002/qua.24890
OpenAlex
W2055526416
Cited by
884
References
6
Citations by year
2026: 70 · 2025: 94 · 2024: 95 · 2023: 133 · 2022: 102 · 2021: 105 · 2020: 112 · 2019: 62 · 2018: 47 · 2017: 36 · 2016: 26 · 2015: 2
FWCI
13.32
Type
journal article
Language
en
Open access
Licence
Locations
International Journal of Quantum Chemistry (publishedVersion, cc-by)
Funders
Deutsche Forschungsgemeinschaft (BE3264/3-1, BE3264/5-1, EXC 1069)
Topics
  • Physical Sciences › Materials Science › Materials Chemistry › Machine Learning in Materials Science
  • Physical Sciences › Physics and Astronomy › Atomic and Molecular Physics, and Optics › Advanced Chemical Physics Studies
  • Physical Sciences › Computer Science › Artificial Intelligence › Neural Networks and Applications
Keywords
neural network potentials, machine learning interatomic potentials, potential energy surface, interatomic potentials

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review