Pith. sign in

REVIEW 3 major objections 5 minor 43 references

Universal Machine Learning Potential for Systems with Reduced Dimensionality

T0 review · 3 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read eSEN reaches DFT-level accuracy from isolated molecules to bulk crystals

desk verdict Useful benchmark dataset and mostly honest comparison, but the headline accuracy claim is conditional on successful relaxations and the ORB-2 pre-filtering is a quiet bias. read the letter →

arxiv 2508.15614 v1 pith:5ZXGG3DI submitted 2025-08-21 cond-mat.mtrl-sci physics.comp-ph

Giulio Benedini , Antoine Loew , Matti Hellstrom , Silvana Botti , Miguel A. L. Marques This is my paper
classification cond-mat.mtrl-sciphysics.comp-ph
keywords universalmachinelearninginteratomicpotentialsdimensionalitytransferability0D-3Dbenchmark0123DdatasetgeometryrelaxationeSENequivariantneuralnetworkDFT-levelaccuracy
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper benchmarks eleven universal machine-learning interatomic potentials (uMLIPs) on a new dataset of 40,000 relaxed compounds spanning four dimensionalities: isolated molecules and clusters (0D), nanowires and nanoribbons (1D), atomic layers and slabs (2D), and bulk crystals (3D). It finds that most models, which are trained predominantly on 3D bulk data, lose accuracy as dimensionality decreases, but a few maintain near-DFT performance everywhere. The best overall model, the equivariant Smooth Energy Network (eSEN), keeps energy errors below 10 meV/atom on more than 75% of all systems and position errors around 0.01–0.02 Å across all dimensionalities. The authors conclude that these top potentials are accurate enough to replace DFT for near-equilibrium geometry relaxation and energy evaluation across the full range from isolated atoms to solids, including mixed-dimensionality systems such as surfaces and interfaces.

What carries the argument

The load-bearing components are the 0123D dataset and the relaxation workflow that turns the benchmark into an end-to-end test. The dataset holds 10,000 PBE-relaxed compounds per dimensionality, generated with dimension-specific strategies—a generative model for 3D candidates and PyXtal for lower-dimensional structures—then pre-relaxed with ORB-2 and refined with DFT. The evaluation protocol starts each uMLIP from the DFT equilibrium geometry, relaxes with the FIRE optimizer, and measures the mean absolute error in atomic positions (after Kabsch alignment) and the energy difference relative to the PBE reference, with failures and fragmentation separately recorded. This design tests whether u

What would settle it

Rerun the benchmark with failures included as errors—assigning a failed relaxation an energy error at least as large as the full structure's energy difference, or a position error equal to the cell size—and check whether eSEN still stays below 10 meV/atom and within 0.02 Å across all dimensionalities. A second check: repeat the relaxations starting from slightly perturbed geometries and see if the reported sub-10 meV/atom errors persist.

Watch

Extended reading notes

Core claim

The central claim is that state-of-the-art universal machine-learning interatomic potentials have reached DFT-level accuracy in geometry relaxation and energy prediction across the entire range of system dimensionality, from isolated molecules to bulk crystals. The paper supports this with a purpose-built benchmark: 10,000 relaxed PBE-quality structures for each dimensionality (the 0123D dataset), constructed to avoid overlap with existing training sets. Forcing each uMLIP to relax these structures from the DFT equilibrium geometry, the authors measure failure rates, optimization steps, energy errors, and position errors. Across all four dimensionalities, the best models—ORB-v2, eqV2, and es

Load-bearing premise

The headline error ranges (0.01–0.02 Å, below 10 meV/atom) are computed only over relaxations that converged without fragmentation; for models like eqV2, 620 of 10,000 0D systems failed and are excluded, so counting failures as errors would shift the rankings and the headline numbers.

Editorial extensions

If this is right

  • If the benchmark numbers hold, eSEN and the top rivals can replace DFT for routine geometry relaxation and single-point energies across molecules, wires, layers, and crystals, at a fraction of the cost.
  • Simulations that couple subsystems of different dimensionality—a molecule on a slab, a nanowire on a support—are no longer forced to mix incompatible levels of theory.
  • The measured degradation from 3D to 0D exposes a training-data bias; models trained with more dimension-balanced data should close the gap, a direct incentive for new dataset construction.
  • The high failure rates of the non-conservative models (eqV2, ORB-3d) in low-dimensional relaxation imply that direct force prediction, unless carefully controlled, undermines practical reliability in the very regimes where DFT replacement is most wanted.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A testable extension would be to fine-tune existing uMLIPs on a fraction of the 0123D dataset and measure how much the dimensionality gap closes; the paper's 'do not train on this' request makes this a controlled experiment for future work.
  • The concentration of failures in force-direct (non-conservative) models suggests that an energy-derived force architecture, or a hybrid that projects forces to a conservative form, may be the most robust path for low-dimensional applications—a hypothesis the paper raises but does not itself pursue.
  • Because the benchmark only samples near-equilibrium geometries, the DFT-replacement claim likely holds for relaxation and static energetics, but extrapolating to reactive pathways, finite-temperature dynamics, or charged defects would require separate validation.
  • For practical users, starting relaxation from the DFT geometry (as done here) may inflate apparent performance relative to starting from a rough or random structure; applying both protocols would quantify this gap.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper introduces a benchmark dataset (0123D) of 10,000 PBE-relaxed compounds per dimensionality (0D, 1D, 2D, 3D) and evaluates 11 universal machine learning interatomic potentials (uMLIPs) by re-relaxing these structures and comparing final energies and Kabsch-aligned atomic positions. The central claims are that the best models—particularly eSEN, with ORB-2 and eqV2 for geometries—reach errors of 0.01–0.02 Å in positions and below 10 meV/atom in energies across all dimensionalities, and are therefore accurate enough to replace DFT for near-equilibrium simulations of mixed-dimensional systems.

Significance. If the claims hold, this is a useful contribution: it provides a consistent PBE-level benchmark spanning dimensionalities, uses a sensible geometry metric, and reports code and data availability. The comparison of 11 uMLIPs, including non-conservative and conservative architectures, is timely and valuable. However, the central quantitative claims rest on error statistics computed only for successfully converged relaxations, which weakens the 'across all dimensionalities' headline. The dataset construction also depends on ORB-2 for pre-relaxation and hull filtering, which may bias rankings. These issues are fixable and do not invalidate the dataset itself, but they require revision before the paper's conclusions can be accepted.

major comments (3)
  1. [§III.B, Table II, Figs. 5–6, §II.D] The headline error metrics are conditional on relaxations that converged without fragmentation. Table II shows that eqV2 fails on 620 of 10,000 0D systems and ORB-3d on 89, yet these models are ranked among the 'best performing models for geometry optimization' in §II.D using error distributions that exclude these failures. A failed relaxation represents an unbounded geometry error, so the ranking and the abstract's '0.01–0.02 Å / below 10 meV/atom across all dimensionalities' overstate unconditional performance. Please report failure counts as a primary metric, or include failures as unrelaxed/infinite error, and revise the conclusions accordingly.
  2. [§III.A] For the 0D–2D subsets, candidate structures were pre-relaxed with ORB-2 and selected by distance to the convex hull computed using ORB-2 energies. ORB-2 is then one of the models evaluated and ranked among the best. This selection may favor structures on which ORB-2 has an advantage, potentially inflating its rank relative to models not used in dataset construction. Please test sensitivity, e.g., by constructing an independent subset with a different selector or by reporting rankings on the unfiltered random structures, or explicitly state this as a limitation.
  3. [§III.A and §IV] The text states that the dataset was constructed to 'minimize potential contamination' with uMLIP training sets, but no deduplication protocol is described. The 0D subset includes molecular structures from the Materials Project and generated clusters, which may overlap with training sets such as SPICE, ANI, or MPtrj. Please provide the exact similarity/removal criteria used and report how many structures were removed at each step. This is load-bearing for the benchmark's validity as an unbiased evaluation.
minor comments (5)
  1. [Fig. 1] 'Clockwise from the top' is ambiguous; please label the subpanels explicitly with 0D/1D/2D/3D or use a legend.
  2. [Table I] The header 'N w Targets' is unclear. Define 'w' (weights?) and expand 'Targets' (E, F, S, D/G) directly in the caption.
  3. [§III.A] For 0D systems, 'random space groups' is confusing because molecules do not have periodic space groups; clarify whether this refers to the simulation cell or to molecular point groups.
  4. [§II.D] The statement 'more than 75% of the energy predictions on 0123D dataset have an error lower than 10 meV/atom' should specify whether this is over all 40,000 systems or per dimensionality, and should note that it applies only to converged relaxations.
  5. [§IV] The data availability section gives a general Alexandria URL rather than a direct link to the 0123D dataset; please provide a specific identifier/path.

Circularity Check

1 steps flagged · score 4.0 of 10

ORB-2 geometry ranking is partially self-selected by using ORB-2 to pre-relax and filter the test set; the eSEN accuracy claim remains independent.

  1. other [Sec. III.A (dataset construction) and Sec. II.D (geometry ranking)]
    "All these initial structures were again pre-relaxed with ORB-2 model [9], and the distance to the convex hull was calculated using the ORB-2 energy... Overall the best performing models with respect to the geometry are ORB-2 and eqV2 followed by eSEN."

    The lower-dimensional test structures are defined through ORB-2: candidates are pre-relaxed with ORB-2 and filtered by ORB-2's hull energy before DFT relaxation. The resulting PBE reference geometries are therefore located in basins of the ORB-2 potential-energy surface. When ORB-2 is then benchmarked by relaxing from those same DFT geometries, its near-one-step convergence and low position errors largely reflect its ability to return to its own pre-selected minima, not an independent measure of transferability. The paper's ranking of ORB-2 among the best geometry models is thus partly an artifact of the dataset-generation rule. This does not affect the separate eSEN energy result, which is independent of ORB-2.

full rationale

The paper is an empirical benchmark rather than a derivation, and most of its numbers are genuinely measured against PBE-DFT references. The central eSEN accuracy claim (75% of energy errors below 10 meV/atom) is not circular: eSEN was not used to construct the 0123D dataset, and its errors are compared to independently DFT-relaxed structures. No fitted parameter is renamed as a prediction, and no uniqueness theorem is imported from the authors' prior work. Self-citations (Alexandria, WBM, the generative model) are used as data/tool sources, not to establish the benchmark outcome. The one significant circularity concern is the ORB-2 feedback loop described above: because the lower-dimensional reference set was pre-relaxed and hull-filtered with ORB-2, ORB-2's top geometry ranking on that set is partly self-fulfilling. In addition, the headline 'across all dimensionalities' error statistics are computed only for relaxations that converged without fragmentation (Sec. III.B; Table II reports up to 620 eqV2 0D failures), so the unconditional accuracy statement is overbroad; however, this is a conditional-reporting issue, not a definitional circularity. Overall, the core claim about eSEN has independent content, but the geometry ranking of ORB-2 is partially compromised by the dataset construction.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The benchmark's validity rests on the PBE-DFT reference being the appropriate common ground truth, on the generative and selection pipeline (PyXtal, the Ref. 39 transformer, ORB-2 hull filter) producing representative low-dimensional chemistry, and on atom-index preservation during relaxation. Hand-set thresholds (force convergence, fragmentation limits, step caps) directly shape the failure and error statistics. No new physical entities are introduced; the 0123D dataset is a measurement resource, not an entity.

free parameters (4)
  • force convergence threshold = 40 meV/A
    Hand-picked stopping criterion for uMLIP relaxation; step counts and failure counts (Table II, Fig. 4) depend on it (Sec III.B).
  • dimensionality fragmentation thresholds = 7.5 A (2D), 12.5 A (1D), 20 A (0D)
    Hand-picked thresholds that determine which relaxed structures are discarded from error metrics; they directly shape the reported position MAE and energy errors (Sec III.B).
  • maximum optimization steps = 15000
    Convergence-failure cutoff; systems exceeding it are labeled unconverged and excluded from error statistics (Sec III.B).
  • force divergence cap = 10000 eV/A
    Labels a relaxation as failed; affects Table II failure counts and survivorship in Figs. 5-6 (Sec III.B).
assumptions (4)
  • domain assumption PBE total energies and relaxed geometries are the correct reference for evaluating uMLIP transferability across dimensionalities.
    The benchmark treats PBE-DFT (VASP, Alexandria-consistent parameters) as ground truth; models trained on hybrid-functional or mixed data inherit functional offsets, a limitation the paper itself acknowledges in Sec I.
  • domain assumption The ORB-2 hull-distance filter selects a representative set of low-dimensional compounds close to thermodynamic stability.
    Sec III.A: for 0D/1D/2D, distance to the convex hull is estimated with ORB-2 energies because no reliable model exists; if these stability estimates are biased, the 0123D test sample is biased and the model ranking is distorted.
  • domain assumption Atom-to-atom correspondence is preserved during relaxation; no atomic permutations occur.
    Sec III.C: the geometry comparison maps uMLIP atoms to PBE reference atoms explicitly under this assumption; permutations would make the Kabsch-aligned MAE meaningless.
  • domain assumption The 3D generative model of Ref. 39 and PyXtal random generation sample chemically meaningful structure space for each subset.
    Sec III.A: 3D candidates come from a generative transformer (Ref. 39, with author overlap), reduced-dimensionality candidates from random PyXtal structures with charge neutrality; coverage of realistic low-dimensional chemistry rests on these generators.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Universal Machine Learning Potential for Systems with Reduced Dimensionality." pith.science (2026). https://pith.science/paper/5ZXGG3DI

@misc{pith2026250815614,
  author       = {Pith},
  title        = {Pith review of: Universal Machine Learning Potential for Systems with Reduced Dimensionality},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5ZXGG3DI}},
  note         = {Machine review of arXiv:2508.15614}
}
read the original abstract

We present a benchmark designed to evaluate the predictive capabilities of universal machine learning interatomic potentials across systems of varying dimensionality. Specifically, our benchmark tests zero- (molecules, atomic clusters, etc.), one- (nanowires, nanoribbons, nanotubes, etc.), two- (atomic layers and slabs) and three-dimensional (bulk materials) compounds. The benchmark reveals that while all tested models demonstrate excellent performance for three-dimensional systems, accuracy degrades progressively for lower-dimensional structures. The best performing models for geometry optimization are orbital version 2, equiformerV2, and the equivariant Smooth Energy Network, with the equivariant Smooth Energy Network also providing the most accurate energies. Our results indicate that the best models yield, on average, errors in the atomic positions in the range of 0.01-0.02 angstrom and errors in the energy below 10~meV/atom across all dimensionalities. These results demonstrate that state-of-the-art universal machine learning interatomic potentials have reached sufficient accuracy to serve as direct replacements for density functional theory calculations, at a small fraction of the computational cost, in simulations spanning the full range from isolated atoms to bulk solids. More significantly, the best performing models already enable efficient simulations of complex systems containing subsystems of mixed dimensionality, opening new possibilities for modeling realistic materials and interfaces.

Figures

Figures reproduced from arXiv: 2508.15614 by the authors.

Figure 2
Figure 2. FIG. 2. Fraction of systems containing a specific element of [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 1
Figure 1. FIG. 1. Scatter plot projection of the 0123D dataset using [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 3
Figure 3. FIG. 3. Distributions of (a) number of atoms, (b) the dis [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: FIG. 4. The box plot distribution of the number of steps [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: FIG. 5. Distribution of energies differences per atom for each [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: FIG. 6. The box plot distribution of the mean absolute error [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

43 extracted references · 22 canonical work pages

  1. [1]

    Behler, Perspective: Machine learning potentials for atomistic simulations, J

    J. Behler, Perspective: Machine learning potentials for atomistic simulations, J. Chem. Phys.145, 170901 (2016)

    2016

  2. [2]

    Schmidt, M

    J. Schmidt, M. R. G. Marques, S. Botti, and M. A. L. Marques, Recent advances and applications of machine learning in solid-state materials science, npj Comput. Mater.5, 83 (2019)

    2019

  3. [3]

    G. Wang, C. Wang, X. Zhang, Z. Li, J. Zhou, and Z. Sun, Machine learning interatomic potential: Bridge the gap between small-scale models and realistic device-scale sim- ulations, iScience27, 109673 (2024)

    work page 2024

  4. [4]

    Chen and S

    C. Chen and S. P. Ong, A universal graph deep learning interatomic potential for the periodic table, Nat. Com- put. Sci.2, 718 (2022)

    work page 2022

  5. [5]

    Y. Park, J. Kim, S. Hwang, and S. Han, Scalable parallel algorithm for graph neural network interatomic poten- tialsinmoleculardynamicssimulations,J.Chem.Theory Comput.20, 4857–4868 (2024)

    work page 2024

  6. [6]

    Batatia, D

    I. Batatia, D. P. Kovács, G. N. C. Simm, C. Ortner, and G. Csányi, MACE: Higher order equivariant message passing neural networks for fast and accurate force fields, arXiv , 2206.07697 (2022)

    arXiv 2022

  7. [7]

    B. Deng, P. Zhong, K. Jun, J. Riebesell, K. Han, C. J. Bartel, and G. Ceder, CHGNet: Pretrained universal neural network potential for charge-informed atomistic modeling, arXiv , 2302.14231 (2023)

    arXiv 2023

  8. [8]

    Barroso-Luque, M

    L. Barroso-Luque, M. Shuaibi, X. Fu, B. M. Wood, M. Dzamba, M. Gao, A. Rizvi, C. L. Zitnick, and Z. W. Ulissi, Open materials 2024 (OMat24) inorganic materi- als dataset and models, arXiv , 2410.12771 (2024)

    arXiv 2024

Show all 43 references
  1. [9]

    Neumann, J

    M. Neumann, J. Gin, B. Rhodes, S. Bennett, Z. Li, H. Choubisa, A. Hussey, and J. Godwin, Orb: A fast, scalable neural network potential, arXiv , 2410.22570 (2024)

    2024 arXiv

  2. [10]

    H. Yang, C. Hu, Y. Zhou, X. Liu, Y. Shi, J. Li, G. Li, Z. Chen, S. Chen, C. Zeni, M. Horton, R. Pinsler, A. Fowler, D. Zügner, T. Xie, J. Smith, L. Sun, Q. Wang, L. Kong, C. Liu, H. Hao, and Z. Lu, Mattersim: A deep learning atomistic model across elements, temperatures and pr...

    2024 arXiv

  3. [11]

    Riebesell, R

    J. Riebesell, R. E. A. Goodall, P. Benner, Y. Chi- ang, B. Deng, A. A. Lee, A. Jain, and K. A. Persson, Matbench discovery – a framework to evaluate machine learning crystal stability predictions, arXiv , 2308.14920 (2023)

    2023 arXiv

  4. [12]

    V. Fung, J. Zhang, E. Juarez, and B. G. Sumpter, Bench- marking graph neural networks for materials chemistry, npj Comput. Mater.7, 84 (2021)

    2021

  5. [13]

    Focassio, L

    B. Focassio, L. P. M. Freitas, and G. R. Schleder, Per- formance assessment of universal machine learning inter- atomic potentials: Challenges and directions for materi- als’ surfaces, ACS Appl. Mater. Interfaces (2024)

    2024

  6. [14]

    H. Yu, M. Giantomassi, G. Materzanini, J. Wang, and G.-M. Rignanese, Systematic assessment of various uni- versal machine-learning interatomic potentials, Mater. Genome Eng. Adv.2, e58 (2024)

    2024

  7. [15]

    Jain, S.P.Ong, G.Hautier, W.Chen, W.D.Richards, S

    A. Jain, S.P.Ong, G.Hautier, W.Chen, W.D.Richards, S. Dacek, S. Cholia, D. Gunter, D. Skinner, G. Ceder, and K. A. Persson, Commentary: The Materials Project: A materials genome approach to accelerating materials innovation, APL Mater.1, 011002 (2013)

    2013

  8. [16]

    Schmidt, T

    J. Schmidt, T. F. Cerqueira, A. H. Romero, A. Loew, F. Jäger, H.-C. Wang, S. Botti, and M. A. Marques, Improving machine-learning models in materials science through large datasets, Mater. Today Phys.48, 101560 (2024)

    2024

  9. [17]

    Devereux, J

    C. Devereux, J. S. Smith, K. K. Huddleston, K. Barros, R. Zubatyuk, O. Isayev, and A. E. Roitberg, Extending the Applicability of the ANI Deep Learning Molecular Potential to Sulfur and Halogens, J. Chem. Theory Com- put.16, 4192 (2020)

    2020

  10. [18]

    Eastman, P

    P. Eastman, P. K. Behara, D. L. Dotson, R. Galvelis, J. E. Herr, J. T. Horton, Y. Mao, J. D. Chodera, B. P. Pritchard, Y.Wang, G.DeFabritiis,andT.E.Markland, Spice, a dataset of drug-like molecules and peptides for training machine learning potentials, Sci. Data10, 11 (2023)

    2023

  11. [19]

    Eastman, B

    P. Eastman, B. P. Pritchard, J. D. Chodera, and T. E. Markland, Nutmeg and SPICE: Models and Data for Biomolecular Machine Learning, J. Chem. Theory Com- put.20, 8583 (2024)

    2024

  12. [20]

    Ganscha, O

    S. Ganscha, O. T. Unke, D. Ahlin, H. Maennel, S. Kashu- bin, and K.-R. Müller, The qcml dataset, quantum chem- istry reference data from 33.5m dft and 14.7b semi- empirical calculations, Sci. Data12, 406 (2025)

    2025

  13. [21]

    A. D. Becke, Density-functional thermochemistry. 3. The role of exact exchange, J. Chem. Phys.98, 5648–5652 (1993)

    1993

  14. [22]

    C. Lee, W. Yang, and R. G. Parr, Development of the colle-salvetti correlation-energy formula into a functional of the electron density, Phys. Rev. B37, 785–789 (1988)

    1988

  15. [23]

    J. P. Perdew, K. Burke, and M. Ernzerhof, Generalized gradient approximation made simple, Phys. Rev. Lett. 77, 3865 (1996)

    1996

  16. [24]

    J. Zeng, D. Zhang, A. Peng, X. Zhang, S. He, Y. Wang, X. Liu, H. Bi, Y. Li, C. Cai, C. Zhang, Y. Du, J.-X. Zhu, P. Mo, Z. Huang, Q. Zeng, S. Shi, X. Qin, Z. Yu, C. Luo, Y.Ding, Y.-P.Liu, R.Shi, Z.Wang, S.L.Bore, J.Chang, Z. Deng, Z. Ding, S. Han, W. Jiang, G. Ke, Z. Liu, D. Lu...

    2025

  17. [25]

    X. Fu, B. M. Wood, L. Barroso-Luque, D. S. Levine, M. Gao, M. Dzamba, and C. L. Zitnick, Learning smooth and expressive interatomic potentials for physical prop- erty prediction, arXiv , 2502.12147 (2025)

    2025 arXiv

  18. [26]

    Bochkarev, Y

    A. Bochkarev, Y. Lysogorskiy, and R. Drautz, Graph atomic cluster expansion for semilocal interactions be- yond equivariant message passing, Phys. Rev. X14, 021036 (2024)

    2024

  19. [27]

    Batatia, P

    I. Batatia, P. Benner, Y. Chiang, A. M. Elena, D. P. Kovács, J. Riebesell, X. R. Advincula, M. Asta, M. Avaylon, W. J. Baldwin, F. Berger, N. Bernstein, A. Bhowmik, S. M. Blau, V. Cărare, J. P. Darby, S. De, F. D. Pia, V. L. Deringer, R. Elijošius, Z. El-Machachi, F. Falcioni,...

    2024 arXiv

  20. [28]

    Rhodes, S

    B. Rhodes, S. Vandenhaute, V. Šimkus, J. Gin, J. God- win, T. Duignan, and M. Neumann, Orb-v3: Atomistic simulation at scale, arXiv , 2504.06231 (2025)

    2025 arXiv

  21. [29]

    J. Kim, J. Kim, J. Kim, J. Lee, Y. Park, Y. Kang, and S. Han, Data-efficient multifidelity training for high- fidelity machine learning interatomic potentials, J. Am. Chem. Soc.147, 1042 (2024)

    2024

  22. [30]

    Sanchez-Gonzalez, J

    A. Sanchez-Gonzalez, J. Godwin, T. Pfaff, R. Ying, J. Leskovec, and P. Battaglia, Learning to simulate com- plex physics with graph networks, inProceedings of the 37th International Conference on Machine Learn- ing, Proceedings of Machine Learning Research, Vol. 119, edited by...

    2020

  23. [31]

    Batzner, A

    S. Batzner, A. Musaelian, L. Sun, M. Geiger, J. P. Mailoa, M. Kornbluth, N. Molinari, T. E. Smidt, and B. Kozinsky, E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials, Nat. Commun.13, 2453 (2022)

    2022

  24. [32]

    Drautz, Atomic cluster expansion for accurate and transferable interatomic potentials, Phys

    R. Drautz, Atomic cluster expansion for accurate and transferable interatomic potentials, Phys. Rev. B99, 014104 (2019)

    2019

  25. [33]

    Zhang, H

    D. Zhang, H. Bi, F.-Z. Dai, W. Jiang, X. Liu, L. Zhang, and H. Wang, Pretraining of attention-based deep learn- ing potential model for molecular simulation, npj Com- put. Mater.10, 94 (2024)

    2024

  26. [34]

    Zhang, X

    D. Zhang, X. Liu, X. Zhang, C. Zhang, C. Cai, H. Bi, Y. Du, X. Qin, A. Peng, J. Huang, B. Li, Y. Shan, J. Zeng, Y. Zhang, S. Liu, Y. Li, J. Chang, X. Wang, S. Zhou, J. Liu, X. Luo, Z. Wang, W. Jiang, J. Wu, Y. Yang, J. Yang, M. Yang, F.-Q. Gong, L. Zhang, M. Shi, F.-Z. Dai, D....

    2024

  27. [35]

    H.-C. Wang, S. Botti, and M. A. L. Marques, Predicting stable crystalline compounds using chemical similarity, Npj Comput. Mater.7, 12 (2021)

    2021

  28. [36]

    C. Zeni, R. Pinsler, D. Zügner, A. Fowler, M. Hor- ton, X. Fu, S. Shysheya, J. Crabbé, L. Sun, J. Smith, B. Nguyen, H. Schulz, S. Lewis, C.-W. Huang, Z. Lu, Y. Zhou, H. Yang, H. Hao, J. Li, R. Tomioka, and T. Xie, MatterGen: A generative model for inorganic materials design, a...

    2024 arXiv

  29. [37]

    Riebesell, H

    J. Riebesell, H. Yang, R. Goodall, and S. G. Baird, Pymatviz: visualization toolkit for mate- rials informatics (2022), 10.5281/zenodo.7486816 - https://github.com/janosh/pymatviz

    2022 doi

  30. [38]

    Merchant, S

    A. Merchant, S. Batzner, S. S. Schoenholz, M. Aykol, G. Cheon, and E. D. Cubuk, Scaling deep learning for materials discovery, Nature624, 80 (2023)

    2023

  31. [39]

    De Breuck, H

    P.-P. De Breuck, H. A. Piracha, G.-M. Rignanese, and M. A. L. Marques, A generative material transformer us- ing wyckoff representation, arXiv , 2501.16051 (2025)

    2025 arXiv

  32. [40]

    Fredericks, K

    S. Fredericks, K. Parrish, D. Sayre, and Q. Zhu, Pyx- tal: A python library for crystal structure generation and symmetry analysis, Comput. Phys. Commun.261, 107810 (2021)

    2021

  33. [41]

    Hjorth Larsen, J

    A. Hjorth Larsen, J. Jørgen Mortensen, J. Blomqvist, I. E. Castelli, R. Christensen, M. Dułak, J. Friis, M. N. Groves, B. Hammer, C. Hargus, E. D. Hermes, P. C. Jennings, P. Bjerre Jensen, J. Kermode, J. R. Kitchin, E. Leonhard Kolsbjerg, J. Kubal, K. Kaasb- jerg, S. Lysgaard,...

    2017

  34. [42]

    Bitzek, P

    E. Bitzek, P. Koskinen, F. Gähler, M. Moseler, and P. Gumbsch, Structural relaxation made simple, Phys. Rev. Lett.97, 170201 (2006)

    2006

  35. [43]

    Kabsch, A solution for the best rotation to relate two sets of vectors, Foundations of Crystallography32, 922 (1976)

    W. Kabsch, A solution for the best rotation to relate two sets of vectors, Foundations of Crystallography32, 922 (1976)

    1976

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.