pith. sign in

arxiv: 2607.02219 · v1 · pith:OZDA4RN7new · submitted 2026-07-02 · ❄️ cond-mat.mtrl-sci · cond-mat.dis-nn· cond-mat.stat-mech

Differentiable inverse design of short-range order in high-entropy alloys: from target sro to target property

Tiancheng Ding , Conrard Giresse Tetsassi Feugmo This is my paper

Pith reviewed 2026-07-03 09:34 UTC · model grok-4.3

classification ❄️ cond-mat.mtrl-sci cond-mat.dis-nncond-mat.stat-mech
keywords high-entropy alloysshort-range orderinverse designgradient-based optimizationmechanical propertiesalloy designdifferentiable methods
0
0 comments X

The pith

Gradient optimization designs SRO in alloys to hit target stiffness

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

The paper introduces a gradient-based method for inverse design of short-range order in high-entropy alloys, allowing direct optimization from target properties to structure. By treating atom occupancy as continuous variables, it enables gradient descent to build structures matching desired SRO patterns, replacing slower random-swap search. A physics-based correction term, extended to multi-element cases, maintains thermodynamic realism in the outputs. A small neural network then predicts mechanical properties from composition and SRO statistics to close the design loop. Tests across nine FCC and BCC alloys captured stiffness shifts from -20% to +57%, with cell-size checks showing that at least 864 atoms are required for reliable results and that the method matched three of four targets within 6% on a real CoCrNi case.

Core claim

By making atom occupancy continuous and applying gradient descent with an extended physics-based correction term, the method generates thermodynamically realistic alloy structures that match target short-range order and produce specified mechanical properties, forming a closed pipeline from target property back to atomic arrangement.

What carries the argument

Continuous atom-occupancy representation optimized by gradient descent, paired with an extended physics-based correction term for thermodynamic realism and a neural-network property predictor.

If this is right

  • The builder matches random-swap accuracy on small systems but runs six times faster and eight times more accurately on 4000-atom systems.
  • Simulation cells need at least 864 atoms to capture the correct direction and magnitude of SRO-driven stiffness changes.
  • The approach scales smoothly to alloys with many elements without added bookkeeping.
  • It reproduced three of four target stiffness values within 6% when checked against real simulations for a cobalt-chromium-nickel alloy.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same continuous-occupancy gradient framework could be retrained on other properties such as yield strength or thermal conductivity to expand the design targets.
  • Literature results based on 108-atom cells may systematically under- or over-estimate SRO effects on properties.
  • Releasing the method as open-source Python code allows direct testing on new alloy families and property combinations.

Load-bearing premise

That extending the physics-based correction term from two-element to many-element alloys sufficiently enforces thermodynamic realism in the optimized structures rather than merely matching numerical SRO targets.

What would settle it

Running independent molecular dynamics simulations on the designed structures for additional alloys beyond the nine tested and checking whether the resulting stiffness values match the pipeline predictions within 6%.

Figures

Figures reproduced from arXiv: 2607.02219 by Conrard Giresse Tetsassi Feugmo, Tiancheng Ding.

Figure 1
Figure 1. Figure 1: ANISRO overview. ANISRO connects any interatomic potential to gradient-based design of atomic short [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Concept of the differentiable SRO-to-structure inverse. Each lattice site carries a logit vector [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Property descriptor and surrogate workflow. Top (blue/green/orange): inference path [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Complete ANISRO pipeline. Blue: forward NIMM extracts effective pair interactions {V (r) ij } and pre￾dicts αeq by Newton iteration. Green: differentiable SRO-to-structure inverse with logit field ℓ, soft occupancy P = σ(ℓ), differentiable α counter, and four-term loss minimised by Adam. Purple: descriptor MLP ϕθ trained on [c, α(1), α(2), σocc]. Orange: closed design loop back-propagates property error th… view at source ↗
Figure 5
Figure 5. Figure 5: Forward NIMM validation against Metropolis Monte Carlo for Cu–Ni. Warren–Cowley [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Loss vs. wall-time for ARMC, gradient, and hybrid methods on Cu [PITH_FULL_IMAGE:figures/full_fig_p011_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Atomic structures of Co–Cr–Ni (4×4×4 FCC, 256 atoms) produced by the gradient inverse (left) and ARMC (right), both targeting α (1) CoCr = −0.38. [001] top-down slab, two atomic layers; atoms coloured by species (Co = orange, Cr = blue, Ni = green). Both methods encode the same Co–Cr hetero-pair enrichment. Structures ex￾ported as LAMMPS data files and rendered directly from atomic coordinates. The hybrid’… view at source ↗
Figure 8
Figure 8. Figure 8: EPI extraction cross-validation on Cu–Ni MEAM. Three routes (three-energy, regression on 60 con [PITH_FULL_IMAGE:figures/full_fig_p013_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: SRO-induced property changes from Table 1. Bars: [PITH_FULL_IMAGE:figures/full_fig_p015_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Atomic structure of the gradient-designed Cr–Fe–Ni configuration ( [PITH_FULL_IMAGE:figures/full_fig_p016_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Cantor alloy (CoNiCrFeMn) with Choi–Lee 2NN-MEAM [11]. [PITH_FULL_IMAGE:figures/full_fig_p018_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Isometric view of the gradient-designed Cantor alloy structure (CoNiCrFeMn, [PITH_FULL_IMAGE:figures/full_fig_p019_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Same Cantor alloy pipeline with MACE-MP-0 via [PITH_FULL_IMAGE:figures/full_fig_p020_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Scaling on Co–Cr–Ni (Choi–Lee 2NN-MEAM [11]) at 256, 864, and 4000 atoms. Target: Metropolis MC [PITH_FULL_IMAGE:figures/full_fig_p021_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Closed-loop property design on Co–Cr–Ni (Choi–Lee 2NN-MEAM [11]). [PITH_FULL_IMAGE:figures/full_fig_p022_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Master summary. (A) ARMC vs. gradient vs. hybrid on Cu50Ni50 at fixed wall time: ARMC wins at 256 atoms, gradient wins beyond 864 atoms. (B) Cross-system SRO-induced ∆C11: Cr–Fe–Ni +57 % stiffening (largest), Cr–Cu–Ni −20 % softening. (C) Scaling envelope: gradient is 6× faster and 8× more accurate than ARMC at 4000 atoms. (D) Closed-loop C11 verification on Co–Cr–Ni: three of four targets within 6 %. 4.1… view at source ↗
read the original abstract

Short-range order (SRO) governs the mechanical response of multi-principal-element alloys, but designing an alloy for a target property usually means solving two disconnected problems: building a structure matching a desired SRO pattern, then separately checking its property, with no shared optimization. This work replaces the standard random-swap search (reverse Monte Carlo) with a gradient-based approach: atom occupancy is treated as continuous rather than fixed, so the whole process can be tuned using gradient descent, the same method used to train neural networks. This builder matches random-swap accuracy on small systems, but is six times faster and eight times more accurate on large 4000-atom systems, and scales smoothly to alloys with many elements without extra bookkeeping. A physics-based correction term, adapted from prior two-element work and extended here to many elements, keeps designed structures thermodynamically realistic rather than just numerically matching the target SRO pattern. A small neural network then predicts mechanical properties directly from composition and SRO statistics, closing the loop from target property back to structure. Tested on nine face-centered-cubic and body-centered-cubic alloys, the pipeline captured SRO-driven stiffness changes from -20% to +57%, and cell-size checks showed at least 864 atoms are needed to get the direction and size of these changes right, since the commonly used 108-atom cells can mislead. Against real simulations for a cobalt-chromium-nickel alloy, the method matched three of four target stiffness values within 6%. The method is released as an open-source Python package, anisro, offering a practical route to gradient-based, property-driven alloy design.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript claims a gradient-based inverse-design pipeline for short-range order (SRO) in high-entropy alloys that treats atom occupancy as continuous, optimizes via gradient descent, employs a neural-network property predictor, and uses an extended physics-based correction term (adapted from two-element prior work) to enforce thermodynamic realism. On nine FCC and BCC alloys the method is reported to recover SRO-driven stiffness changes of -20% to +57%, requires cells of at least 864 atoms, and matches three of four target stiffness values within 6% for CoCrNi; the code is released as the open-source package anisro.

Significance. If the thermodynamic-realism claim holds, the work supplies a scalable, differentiable route from target property to SRO structure that is substantially faster and more accurate than reverse Monte Carlo on large cells and removes the need for disconnected forward and inverse steps. The open-source release is a concrete strength that supports reproducibility.

major comments (2)
  1. [Abstract] Abstract (description of the correction term): the claim that the physics-based correction term, when extended to many elements, keeps structures 'thermodynamically realistic rather than just numerically matching the target SRO pattern' is load-bearing for the central assertion that the pipeline yields property-relevant SRO. No explicit multi-element energy functional, derivation, or validation against independent Monte Carlo pair-correlation thermodynamics or formation energies is supplied; without this the extension risks reducing to a soft constraint on Warren-Cowley parameters.
  2. [Abstract] Abstract (performance claims): the statements of 'six times faster and eight times more accurate on large 4000-atom systems' and 'at least 864 atoms are needed' are presented without error bars, dataset descriptions, or cross-validation details. These metrics are central to the superiority claim over random-swap search and must be supported by explicit methods and statistics.
minor comments (1)
  1. The abstract reports concrete numerical matches (6% on CoCrNi) but supplies no description of the neural-network architecture, training set size, or loss function; these details belong in the main text for reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their careful reading and constructive comments on our manuscript. We address each major comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract (description of the correction term): the claim that the physics-based correction term, when extended to many elements, keeps structures 'thermodynamically realistic rather than just numerically matching the target SRO pattern' is load-bearing for the central assertion that the pipeline yields property-relevant SRO. No explicit multi-element energy functional, derivation, or validation against independent Monte Carlo pair-correlation thermodynamics or formation energies is supplied; without this the extension risks reducing to a soft constraint on Warren-Cowley parameters.

    Authors: We agree that the abstract claim requires explicit supporting material. The correction term is an extension of the two-element formulation, but the manuscript does not currently provide the multi-element energy functional, its derivation, or direct validation against independent Monte Carlo thermodynamics. In the revised manuscript we will add a dedicated subsection in Methods with the explicit functional form and a new figure comparing pair correlations and formation energies of the optimized structures to independent Monte Carlo results on the same systems. revision: yes

  2. Referee: [Abstract] Abstract (performance claims): the statements of 'six times faster and eight times more accurate on large 4000-atom systems' and 'at least 864 atoms are needed' are presented without error bars, dataset descriptions, or cross-validation details. These metrics are central to the superiority claim over random-swap search and must be supported by explicit methods and statistics.

    Authors: We acknowledge that the performance statements in the abstract are presented without the requested statistical support. In the revised manuscript we will expand the Results section to report error bars from repeated runs, describe the full dataset (alloys and cell sizes tested), and detail the cross-validation procedure used to quantify the speedup and accuracy gains on 4000-atom systems as well as the cell-size convergence analysis. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation remains self-contained

full rationale

The pipeline consists of a gradient-based optimizer on continuous atom occupancies, a correction term extended from prior two-element work (explicitly stated as adapted and extended in the present manuscript), and a separately trained neural network for property prediction from composition and SRO statistics. No equation or step reduces by construction to a fitted input renamed as output, nor does any load-bearing premise collapse to an unverified self-citation chain. The central loop from target property to structure is externally falsifiable via the reported comparisons to Monte Carlo and experimental stiffness values, satisfying the criteria for independent content.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

The central claim rests on continuous relaxation of discrete atom occupancies and on a trained neural-network surrogate; both introduce fitted elements whose values are not supplied in the abstract.

free parameters (2)
  • neural-network weights
    Trained to map composition and SRO statistics to mechanical properties; values not reported in abstract.
  • correction-term coefficients
    Adapted from prior two-element work and extended to multi-element alloys; specific values not given.
axioms (2)
  • domain assumption Continuous relaxation of atom occupancy remains valid for producing thermodynamically realistic structures when combined with the physics correction.
    Enables gradient descent; invoked to replace discrete random-swap search.
  • domain assumption The neural network provides a sufficiently accurate forward map from SRO to stiffness for the inverse-design loop to be useful.
    Closes the property-to-structure optimization; no independent validation details in abstract.

pith-pipeline@v0.9.1-grok · 5852 in / 1517 out tokens · 36479 ms · 2026-07-03T09:34:25.247006+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

48 extracted references · 45 canonical work pages · 2 internal anchors

  1. [1]

    Allen, Tatiana Lopez-Guevara, Kimberly Stachenfeld, Alvaro Sanchez-Gonzalez, Peter Battaglia, Jessica Hamrick, and Tobias Pfaff

    Kelsey R. Allen, Tatiana Lopez-Guevara, Kimberly Stachenfeld, Alvaro Sanchez-Gonzalez, Peter Battaglia, Jessica Hamrick, and Tobias Pfaff. Physical design using differentiable learned simulators, 2022

    2022

  2. [2]

    Mu \ n oz, J

    Mattias ngqvist, William A. Mu \ n oz, J. Magnus Rahm, Erik Fransson, C \'e line Durniak, Piotr Rozyczko, Thomas H. Rod, and Paul Erhart. ICET -- a python library for constructing and sampling alloy cluster expansions. Advanced Theory and Simulations, 2 0 (7): 0 1900015, 2019. doi:10.1002/adts.201900015

    work page doi:10.1002/adts.201900015 2019

  3. [3]

    Woodward, S

    Edwin Antillon, C. Woodward, S. I. Rao, B. Akdim, and T. A. Parthasarathy. Chemical short range order strengthening in a model FCC high entropy alloy. Acta Materialia, 190: 0 29--42, 2020. doi:10.1016/j.actamat.2020.02.041

    work page doi:10.1016/j.actamat.2020.02.041 2020

  4. [4]

    Kov \'a cs, Gregor N

    Ilyes Batatia, D \'a vid P. Kov \'a cs, Gregor N. C. Simm, Christoph Ortner, and G \'a bor Cs \'a nyi. MACE : Higher order equivariant message passing neural networks for fast and accurate force fields. In Advances in Neural Information Processing Systems (NeurIPS), volume 35, 2022

    2022

  5. [5]

    Mailoa, Mordechai Kornbluth, Nicola Molinari, Tess E

    Simon Batzner, Albert Musaelian, Lixin Sun, Mario Geiger, Jonathan P. Mailoa, Mordechai Kornbluth, Nicola Molinari, Tess E. Smidt, and Boris Kozinsky. E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials. Nature Communications, 13: 0 2453, 2022. doi:10.1038/s41467-022-29939-5

    work page doi:10.1038/s41467-022-29939-5 2022

  6. [6]

    Structural relaxation made simple

    Erik Bitzek, Pekka Koskinen, Franz G \"a hler, Michael Moseler, and Peter Gumbsch. Structural relaxation made simple. Physical Review Letters, 97: 0 170201, 2006. doi:10.1103/PhysRevLett.97.170201

    work page doi:10.1103/physrevlett.97.170201 2006

  7. [7]

    Cantor, I

    B. Cantor, I. T. H. Chang, P. Knight, and A. J. B. Vincent. Microstructural development in equiatomic multicomponent alloys. Materials Science and Engineering: A, 375--377: 0 213--218, 2004. doi:10.1016/j.msea.2003.10.257

    work page doi:10.1016/j.msea.2003.10.257 2004

  8. [8]

    Mechanical properties of high-entropy alloys with compositional disorder at small scales

    Penghui Cao. Mechanical properties of high-entropy alloys with compositional disorder at small scales. npj Computational Materials, 7: 0 166, 2021. doi:10.1038/s41524-021-00647-y

    work page doi:10.1038/s41524-021-00647-y 2021

  9. [9]

    Direct observation of chemical short-range order in a medium-entropy alloy

    Xuefei Chen, Qin Wang, Zhiying Cheng, Mingli Zhu, Hao Zhou, Ping Jiang, Lingling Zhou, Qiqige Xue, Fuping Yuan, Jing Zhu, Xiaolei Wu, and En Ma. Direct observation of chemical short-range order in a medium-entropy alloy. Nature, 592: 0 712--716, 2021. doi:10.1038/s41586-021-03428-z

    work page doi:10.1038/s41586-021-03428-z 2021

  10. [10]

    Structure motif of chemical short-range order in a medium-entropy alloy

    Xuefei Chen, Fuping Yuan, Hao Zhou, and Xiaolei Wu. Structure motif of chemical short-range order in a medium-entropy alloy. Materials Research Letters, 10: 0 149--155, 2022. doi:10.1080/21663831.2022.2029607

    work page doi:10.1080/21663831.2022.2029607 2022

  11. [11]

    Understanding the physical metallurgy of the CoCrFeMnNi high-entropy alloy: an atomistic simulation study

    Won-Mi Choi, Yong Hee Jo, Seok Su Sohn, Sunghak Lee, and Byeong-Joo Lee. Understanding the physical metallurgy of the CoCrFeMnNi high-entropy alloy: an atomistic simulation study. npj Computational Materials, 4: 0 1, 2018. doi:10.1038/s41524-018-0060-9

    work page doi:10.1038/s41524-018-0060-9 2018

  12. [12]

    J. M. Cowley. An approximate theory of order in alloys. Physical Review, 77: 0 669--675, 1950. doi:10.1103/PhysRev.77.669

    work page doi:10.1103/physrev.77.669 1950

  13. [13]

    Bartel, and Gerbrand Ceder

    Bowen Deng, Peichen Zhong, KyuJung Jun, Janosh Riebesell, Kevin Han, Christopher J. Bartel, and Gerbrand Ceder. CHGNet as a pretrained universal neural network potential for charge-informed atomistic modelling. Nature Machine Intelligence, 5: 0 1031--1041, 2023. doi:10.1038/s42256-023-00716-3

    work page doi:10.1038/s42256-023-00716-3 2023

  14. [14]

    Jun Ding, Qin Yu, Mark Asta, and Robert O. Ritchie. Tunable stacking fault energies by tailoring local chemical order in CrCoNi medium-entropy alloys. Proceedings of the National Academy of Sciences, 115: 0 8919--8924, 2018. doi:10.1073/pnas.1808660115

    work page doi:10.1073/pnas.1808660115 2018

  15. [15]

    Torchmd: A deep learning framework for molecular simulations

    Stefan Doerr, Maciej Majewski, Adrià Pérez, Andreas Krämer, Cecilia Clementi, Frank Noe, Toni Giorgino, and Gianni De Fabritiis. Torchmd: A deep learning framework for molecular simulations. Journal of Chemical Theory and Computation, 17: 0 2355--2363, 2021. doi:10.1021/acs.jctc.0c01343

    work page doi:10.1021/acs.jctc.0c01343 2021

  16. [16]

    Ant \' o nio Fern \'a ndez-Caballero, Marcel H. F. Sluiter, Jakub S. Wrobel, Duc Nguyen-Manh, Sergei L. Dudarev, and Paul M. Mummery. Short-range order strengthening in BCC transition-metal alloys. Physical Review Materials, 4: 0 023403, 2020. doi:10.1103/PhysRevMaterials.4.023403

    work page doi:10.1103/physrevmaterials.4.023403 2020

  17. [17]

    E. P. George, D. Raabe, and R. O. Ritchie. High-entropy alloys. Nature Reviews Materials, 4: 0 515--534, 2019. doi:10.1038/s41578-019-0121-4

    work page doi:10.1038/s41578-019-0121-4 2019

  18. [18]

    Goodrich, Ella M

    Carl P. Goodrich, Ella M. King, Samuel S. Schoenholz, Ekin D. Cubuk, and Michael P. Brenner. Designing self-assembling kinetics with differentiable statistical physics models. Proceedings of the National Academy of Sciences, 118 0 (10): 0 e2024083118, 2021. doi:10.1073/pnas.2024083118

    work page doi:10.1073/pnas.2024083118 2021

  19. [19]

    Capturing short-range order in high-entropy alloys with machine learning potentials

    Ying Han et al. Capturing short-range order in high-entropy alloys with machine learning potentials. npj Computational Materials, 11: 0 22, 2025. doi:10.1038/s41524-025-01722-2

    work page doi:10.1038/s41524-025-01722-2 2025

  20. [20]

    Uberuaga

    Chao Jiang and Blas P. Uberuaga. Efficient ab initio modeling of random multicomponent alloys. Physical Review Letters, 116: 0 105501, 2016. doi:10.1103/PhysRevLett.116.105501

    work page doi:10.1103/physrevlett.116.105501 2016

  21. [21]

    Adam: A Method for Stochastic Optimization

    Diederik P. Kingma and Jimmy Ba. Adam: A method for stochastic optimization. In International Conference on Learning Representations (ICLR), 2015. URL https://arxiv.org/abs/1412.6980

    work page internal anchor Pith review Pith/arXiv arXiv 2015

  22. [22]

    Development and application of a Ni-Ti interatomic potential with high predictive accuracy of the martensitic phase transition

    Wan-Sik Ko, Blazej Grabowski, and J \"o rg Neugebauer. Development and application of a Ni-Ti interatomic potential with high predictive accuracy of the martensitic phase transition. Physical Review B, 92: 0 134107, 2021. doi:10.1103/PhysRevB.92.134107

    work page doi:10.1103/physrevb.92.134107 2021

  23. [23]

    Guillaume Laplanche, Pascal Gadaud, Moritz H \"a rtelt, and Easo P. George. Single crystal elastic constants of equiatomic CrCoNi medium-entropy alloy. Materials Science and Engineering: A, 823: 0 141762, 2021. doi:10.1016/j.msea.2021.141762

    work page doi:10.1016/j.msea.2021.141762 2021

  24. [24]

    Byeong-Joo Lee and Michael I. Baskes. Second nearest-neighbor modified embedded-atom-method potential. Physical Review B, 62: 0 8564--8567, 2000. doi:10.1103/PhysRevB.62.8564

    work page doi:10.1103/physrevb.62.8564 2000

  25. [25]

    Q.-J. Li, H. Sheng, and E. Ma. Strengthening in multi-principal element alloys with local-chemical-order roughened dislocation pathways. Nature Communications, 10: 0 3563, 2019. doi:10.1038/s41467-019-11464-7

    work page doi:10.1038/s41467-019-11464-7 2019

  26. [26]

    Machine learning modeling of high entropy alloy: the role of short-range order

    Xianglin Liu, Jiaxin Zhang, and Yang Wang. Machine learning modeling of high entropy alloy: the role of short-range order. arXiv preprint arXiv:1906.10889, 2019

    work page arXiv 1906

  27. [27]

    Metropolis, A

    N. Metropolis, A. W. Rosenbluth, M. N. Rosenbluth, A. H. Teller, and E. Teller. Equation of state calculations by fast computing machines. The Journal of Chemical Physics, 21: 0 1087--1092, 1953. doi:10.1063/1.1699114

    work page doi:10.1063/1.1699114 1953

  28. [28]

    D. B. Miracle and O. N. Senkov. A critical review of high entropy alloys and related concepts. Acta Materialia, 122: 0 448--511, 2017. doi:10.1016/j.actamat.2016.08.081

    work page doi:10.1016/j.actamat.2016.08.081 2017

  29. [29]

    Owen, Mordechai Kornbluth, and Boris Kozinsky

    Albert Musaelian, Simon Batzner, Anders Johansson, Lixin Sun, Cameron J. Owen, Mordechai Kornbluth, and Boris Kozinsky. Learning local equivariant representations for large-scale atomistic dynamics. Nature Communications, 14: 0 579, 2023. doi:10.1038/s41467-023-36329-y

    work page doi:10.1038/s41467-023-36329-y 2023

  30. [30]

    PyTorch: An Imperative Style, High-Performance Deep Learning Library

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, et al. PyTorch : An imperative style, high-performance deep learning library. In Advances in Neural Information Processing Systems, volume 32, 2019. URL https://arxiv.org/abs/1912.01703

    work page internal anchor Pith review Pith/arXiv arXiv 2019

  31. [32]

    Youfu Rao and William A. Curtin. Analytical models of short-range order in FCC and BCC alloys. Acta Materialia, 226: 0 117621, 2022 b . doi:10.1016/j.actamat.2022.117621

    work page doi:10.1016/j.actamat.2022.117621 2022

  32. [33]

    J. M. Sanchez, F. Ducastelle, and D. Gratias. Generalized cluster description of multicomponent systems. Physica A, 128 0 (1--2): 0 334--350, 1984. doi:10.1016/0378-4371(84)90096-7

    work page doi:10.1016/0378-4371(84)90096-7 1984

  33. [34]

    Schneeweiss, M

    O. Schneeweiss, M. Fri\' a k, M. Pola s kov\' a , M. Sob, D. Vojt e ch, F. Zelenka, and J. Ve s t\' a l. Magnetic properties of the CrMnFeCoNi high-entropy alloy. Physical Review B, 96: 0 014437, 2017. doi:10.1103/PhysRevB.96.014437

    work page doi:10.1103/physrevb.96.014437 2017

  34. [35]

    Schoenholz and Ekin D

    Samuel S. Schoenholz and Ekin D. Cubuk. JAX MD : A framework for differentiable physics. In Advances in Neural Information Processing Systems (NeurIPS), volume 33, 2020

    2020

  35. [36]

    Chemical-motif characterization of short-range order with E(3) -equivariant graph neural networks

    Killian Sheriff, Yifan Cao, and Rodrigo Freitas. Chemical-motif characterization of short-range order with E(3) -equivariant graph neural networks. npj Computational Materials, 10: 0 215, 2024 a . doi:10.1038/s41524-024-01393-5

    work page doi:10.1038/s41524-024-01393-5 2024

  36. [37]

    Quantifying chemical short-range order in metallic alloys

    Killian Sheriff, Yifan Cao, Tess Smidt, and Rodrigo Freitas. Quantifying chemical short-range order in metallic alloys. Proceedings of the National Academy of Sciences, 121: 0 e2322962121, 2024 b . doi:10.1073/pnas.2322962121

    work page doi:10.1073/pnas.2322962121 2024

  37. [38]

    A. Tamm, A. Aabloo, M. Klintenberg, M. Stocks, and A. Caro. Atomic-scale properties of Ni-based FCC ternary, and quaternary alloys. Acta Materialia, 99: 0 307--312, 2015. doi:10.1016/j.actamat.2015.08.015

    work page doi:10.1016/j.actamat.2015.08.015 2015

  38. [39]

    van de Walle, M

    A. van de Walle, M. Asta, and G. Ceder. The alloy theoretic automated toolkit: A user guide. CALPHAD, 26 0 (4): 0 539--553, 2002. doi:10.1016/S0364-5916(02)80006-2

    work page doi:10.1016/s0364-5916(02)80006-2 2002

  39. [40]

    Flynn Walsh, Mark Asta, and Robert O. Ritchie. Magnetically driven short-range order can explain anomalous measurements in CrCoNi . Proceedings of the National Academy of Sciences, 118 0 (13): 0 e2020540118, 2021. doi:10.1073/pnas.2020540118

    work page doi:10.1073/pnas.2020540118 2021

  40. [41]

    Ritchie, Andrew M

    Flynn Walsh, Jun Zhang, Robert O. Ritchie, Andrew M. Minor, and Mark Asta. Ubiquitous short-range order in multi-principal element alloys. Nature Communications, 15: 0 6602, 2024. doi:10.1038/s41467-024-49606-1

    work page doi:10.1038/s41467-024-49606-1 2024

  41. [42]

    B. E. Warren. X-ray diffraction study of the structure of alloys. Physical Review, 78: 0 821, 1950. doi:10.1103/PhysRev.78.821

    work page doi:10.1103/physrev.78.821 1950

  42. [43]

    Short-range ordering and its effects on mechanical properties of high-entropy alloys

    Yuan Wu, Fei Zhang, Xiaoyuan Yuan, Hailong Huang, Xiaocan Wen, Yihan Wang, Mengyuan Zhang, Honghui Wu, Xiongjun Liu, Hui Wang, Suihe Jiang, and Zhaoping Lu. Short-range ordering and its effects on mechanical properties of high-entropy alloys. Journal of Materials Science & Technology, 62: 0 214--220, 2021 a . doi:10.1016/j.jmst.2020.06.018

    work page doi:10.1016/j.jmst.2020.06.018 2021

  43. [44]

    Pharr, and Easo P

    Zhenggang Wu, Hongbin Bei, George M. Pharr, and Easo P. George. Recovery, recrystallization, grain growth and phase stability of a family of FCC -structured multi-component equiatomic solid solution alloys. Intermetallics, 46: 0 131--140, 2014. doi:10.1016/j.intermet.2013.10.024

    work page doi:10.1016/j.intermet.2013.10.024 2014

  44. [45]

    Deformation mechanisms of high-entropy alloys revealed by in-situ studies

    Zhengxiong Wu, Yongjie Zhao, Haibo Yang, Yu Song, Jinghua Qin, and En Ma. Deformation mechanisms of high-entropy alloys revealed by in-situ studies. Acta Materialia, 215: 0 117058, 2021 b . doi:10.1016/j.actamat.2021.117058

    work page doi:10.1016/j.actamat.2021.117058 2021

  45. [46]

    Yeh, S.-K

    J.-W. Yeh, S.-K. Chen, S.-J. Lin, J.-Y. Gan, T.-S. Chin, T.-T. Shun, C.-H. Tsau, and S.-Y. Chang. Nanostructured high-entropy alloys with multiple principal elements: Novel alloy design concepts and outcomes. Advanced Engineering Materials, 6: 0 299--303, 2004. doi:10.1002/adem.200300567

    work page doi:10.1002/adem.200300567 2004

  46. [47]

    Ritchie, and Andrew M

    Ruopeng Zhang, Shiteng Zhao, Jun Ding, Yan Chong, Tao Jia, Colin Ophus, Mark Asta, Robert O. Ritchie, and Andrew M. Minor. Short-range order and its impact on the CrCoNi medium-entropy alloy. Nature, 581: 0 283--287, 2020. doi:10.1038/s41586-020-2275-z

    work page doi:10.1038/s41586-020-2275-z 2020

  47. [48]

    Atomic-scale evidence of chemical short-range order in CrCoNi medium-entropy alloy

    Lingling Zhou, Qin Wang, Jian Wang, Xuefei Chen, Ping Jiang, Hao Zhou, Fuping Yuan, Xiaolei Wu, Zhiying Cheng, and En Ma. Atomic-scale evidence of chemical short-range order in CrCoNi medium-entropy alloy. Acta Materialia, 224: 0 117490, 2022. doi:10.1016/j.actamat.2021.117490

    work page doi:10.1016/j.actamat.2021.117490 2022

  48. [49]

    Zunger, S.-H

    A. Zunger, S.-H. Wei, L. G. Ferreira, and J. E. Bernard. Special quasirandom structures. Physical Review Letters, 65: 0 353--356, 1990. doi:10.1103/PhysRevLett.65.353

    work page doi:10.1103/physrevlett.65.353 1990