Pith. sign in

REVIEW 5 major objections 4 minor 32 references

Physics-Informed Machine Learning for Refractory Alloy Design

T0 review · 5 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read A physics-informed machine learning model predicts eight mechanical properties of refractory complex concentrated alloys with R² 0.89–0.97 and 100% Born stability compliance on held-out tests.

desk verdict Solid regression work undercut by an internal contradiction in the headline design rule and an overclaimed screening result; worth refereeing after major revision. read the letter →

arxiv 2608.03805 v1 pith:3SODRCSD submitted 2026-08-04 cond-mat.mtrl-sci

Blaise Awola Ayirizia , Bimal K C , Jorge A. Munoz San Martin This is my paper · ORCID
classification cond-mat.mtrl-sci PACS 81.05.Bx71.15.Mb
keywords refractorycomplexconcentratedalloysphysics-informedmachinelearningquantilegradientboostingBornstabilitycriteriaelasticconstantscompositionalscreeningvalenceelectronconcentrationalloydesign
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

The paper argues that combining quantile gradient boosting with Born stability checks—applied to 393 DFT-computed BCC alloys—produces a model accurate enough to screen the roughly 10^15 possible refractory complex concentrated alloy compositions. It claims the model predicts six elastic and hardness properties directly, derives two more from them, and satisfies all mechanical stability criteria on every held-out prediction. Screening then identifies MoReW as the top shear-resistant candidate (predicted C44 = 139 GPa) and yields quantitative design rules: VEC between 6.0 and 6.4, atomic size mismatch below 10%, and Re–Mo–Cr ternary systems maximize shear rigidity. If these claims hold, expensive experimental or DFT searches can be replaced by fast, risk-aware screening with a short list of synthesis-ready candidates.

What carries the argument

The key machinery is quantile gradient boosting regression (QGBR), an ensemble of regression trees that estimates conditional quantiles at 0.05, 0.50, and 0.95, giving both median predictions and 90% prediction intervals. The physics enters through 13 features including valence electron concentration (VEC) and atomic size mismatch (δ), through deriving E and ν from predicted B and G via standard elasticity relations rather than fitting them independently, and through post-prediction verification of the Born stability criteria for cubic crystals. The model does not enforce the Born criteria in its loss function; it instead relies on a fully Born-stable DFT training set and smooth composition–

What would settle it

Measure the elastic constants of arc-melted MoReW by resonant ultrasound spectroscopy and compare C44 to the predicted 139 GPa; a discrepancy well beyond the 8–12 GPa DFT/experiment scatter would undermine the screening. Also compute or measure C44 for a VEC ≈ 6.4 alloy without rhenium (e.g., CrMo3) to test whether the predicted shear penalty is real.

Watch

Extended reading notes

Core claim

The central claim is that a quantile gradient boosting model, trained on 13 composition-derived descriptors from 393 DFT-computed single-phase BCC alloys, can predict bulk modulus, shear modulus, Vickers hardness, and elastic constants C11, C12, and C44 with test-set R² values of 0.89–0.97, while Young's modulus and Poisson's ratio derived from the predicted B and G remain thermodynamically consistent. All 59 held-out test predictions satisfy the Born stability criteria (C11–C12 > 0, C44 > 0, C11 + 2C12 > 0), with no violations. Compositional screening in VEC–δ space reveals a high-shear-rigidity zone near VEC 5.8–6.4 and δ 3–9%, and elemental analysis of the top-ranked candidates shows rhen

Load-bearing premise

The design rules and top-candidate ranking assume the 59 held-out alloys represent the trillion-scale BCC composition space; if that sample is not representative, the rules may not generalize beyond this dataset.

Editorial extensions

If this is right

  • MoReW, already synthesized as a single-phase BCC alloy, now has a quantitative target: measuring C44 near the predicted 139 GPa would directly confirm the screening pipeline.
  • The proposed design rules (VEC = 6.0–6.4, δ < 10%, Re–Mo–Cr ternary preference) give experimentalists a narrow composition window to explore instead of the full trillion-scale space.
  • Re-free alternatives identified in the paper, such as CrMo3, MoW, and CrMoW ternaries, retain roughly 88–92% of the top C44 at much lower cost, offering practical substitutes where rhenium is too expensive.
  • The 90% prediction intervals allow risk-stratified screening: high-confidence predictions can proceed directly to synthesis, while wide-interval candidates can be sent for DFT refinement first.
  • The same workflow—compositional descriptors, quantile boosting, and Born checks—can be transferred to other alloy families and properties, provided a matching DFT or experimental dataset exists.

Reading between the lines

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

  • The 100% Born compliance is observed within the sampled composition window, not proven as a general guarantee; a stress test using out-of-distribution compositions near the BCC/FCC boundary or with δ > 10% would show whether the model's stability holds beyond its training region.
  • Because rhenium's dominance in the top ten is a correlation within 393 DFT alloys, an experimental comparison of MoReW against a Re-free mimic (e.g., CrMo3) would test whether the predicted ~8% shear penalty for removing rhenium is real and persists at service temperatures.
  • The VEC 6.0–6.4 optimum sits near the BCC/FCC phase boundary (VEC ≈ 6.87), so the highest-C44 candidates may be elastically stable yet thermodynamically metastable; pairing this ML screen with CALPHAD or formation-energy calculations would sharpen the design rules.
  • A natural next step is active learning: use the model's prediction intervals to select the most uncertain VEC–δ regions for new DFT calculations, retrain, and see whether the VEC 6.0–6.4 window shifts as the dataset grows beyond 393 alloys.
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

5 major / 4 minor

Summary. The manuscript presents a quantile gradient boosting regression framework, trained on 393 DFT-computed BCC refractory complex concentrated alloys from a 10-element space, to predict six elastic properties (B, G, Hv, C11, C12, C44) and derive E and ν from B and G. The authors report R² = 0.89–0.97 on a 59-alloy test set, claim 100% Born stability satisfaction on test predictions, and derive screening design rules: Re appears in 100% of top-10 alloys, optimal VEC = 6.0–6.4, δ < 10%, and Re–Mo–Cr ternary systems maximize shear rigidity. The top candidate is MoReW with predicted C44 = 139 GPa. The screening and design-rule conclusions are drawn from the 59 test alloys in Section 3.5.

Significance. If the claims held, the framework could be a useful screening tool for refractory CCAs. The paper has genuine strengths: it uses a public DFT dataset, reports learning curves and cross-validation, attempts experimental validation on 18 alloys, and provides quantile-based uncertainty intervals. However, the headline UQ and design-rule claims are compromised by internal inconsistencies and overstatement. The 'Re appears in 100% of top-10' rule is arithmetically contradicted by the paper's own Table 3.3, and the 'physics-informed constraints' are post-hoc diagnostics rather than enforcement. The screening is performed on the 59 test alloys, not on the trillion-scale composition space claimed in the abstract. These issues affect the central contributions, so the paper needs substantial revision before its conclusions can be accepted.

major comments (5)
  1. [Abstract; §3.5; Conclusions] The statement 'rhenium appears in 100% of top-performing alloys' is internally contradicted by Table 3.3, which lists the five highest-ranked alloys by predicted C44: MoReW, CrMo3, Cr2ReV, CrReV, CrReW2. CrMo3 contains no Re. Since the top-5 are necessarily part of the top-10, at most 9 of the top-10 can contain Re. Either the elemental analysis was performed on a different, unreported candidate list, or the frequency is miscalculated. This invalidates the design rule 'Re is essential' and the associated Re–Mo–Cr ternary recommendation. The authors must provide the actual top-10 list and recompute the frequencies.
  2. [Table 3.1; Abstract; Introduction] The paper repeatedly claims 'well-calibrated 90% prediction intervals with 78–90% coverage' for the framework. Table 3.1 reports PICP for ν (Poisson's ratio) as 1.7%. Since ν is one of the eight properties the paper says it predicts, this single value contradicts the 78–90% coverage claim. The footnote 'PICP for ν is not directly comparable due to non-Gaussian distribution' is not sufficient: either derive meaningful intervals for ν or restrict the UQ claim to the six directly modeled properties and report the ν coverage honestly as a failure.
  3. [§2.6.2; §3.3] The abstract and conclusions state that physics-informed Born stability constraints ensure '100% compliance' on test predictions. However, §2.6.2 says compliance was 'evaluated post-prediction' and §3.3 says the model achieves perfect compliance 'without explicit constraint enforcement.' Thus the 100% figure is a post-hoc observation on 59 test points, not evidence that constraints are enforced during training or applied during prediction. This distinction matters because, as the authors note, all 393 training alloys are Born-stable; the model may simply be interpolating. Please reframe the claim as a diagnostic rather than a property of the framework.
  4. [§3.5; Conclusions] The compositional screening and design rules are derived exclusively from the 59 test alloys. Figure 3.3 and the elemental frequencies are summaries of this small test set, not a screening of the trillion-scale composition space. The conclusion that the framework 'enables screening of 10^15 possible compositions' is not demonstrated. To support this claim, the authors need to apply the model to unseen compositions outside the 393-alloy dataset and describe how candidates are selected. As written, the design rules are descriptive statistics of a handful of already-computed alloys.
  5. [§3.4; §4.6] MoReW is presented as the top candidate with predicted C44 = 139 GPa, but the paper never reports the actual DFT C44 for MoReW from the test set. Since MoReW is in the test set, its true value is available and should be compared with the prediction. Without this comparison, the 'discovery' of MoReW as a top candidate could be an artifact of ranking noisy test predictions; the 90% PI width is approximately 40 GPa, which is substantial. Please report the actual C44 and the rank of MoReW (and the other top-5 candidates) in the test set.
minor comments (4)
  1. [§2.6] The heading 'Physics-Informed Constraints' is misleading when the criteria are only evaluated after prediction. Consider renaming to 'Post-hoc physics validation' to match the actual method.
  2. [§4.8 vs Table 3.3] CrMo3 is listed with C44 = 133.22 GPa in Table 3.3 but as 128 GPa in §4.8 ('CrMo3 (C44 = 128 GPa, 8% reduction)'). Reconcile this inconsistency.
  3. [§3.5] The design rule states 'δ < 10%', but the high-C44 region in Figure 3.3 is described as δ = 3–9%. Clarify whether the recommended δ range is the narrower 3–9% or the broader <10%.
  4. [Table 3.1] Use ν instead of µ for Poisson's ratio for consistency with the text. Also, the footnote about ν PICP should be replaced with a substantive explanation of why coverage fails, rather than dismissing it as 'not directly comparable.'

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity—the ML pipeline predicts held-out properties from independent DFT data; screening rules are post-hoc summaries, not inputs.

full rationale

The paper's derivation chain is a standard supervised ML workflow: 393 DFT alloys are split into training (334) and test (59) sets; models predict six elastic properties on the held-out test set; derived properties E and ν are computed from predicted B and G using standard elasticity relations; Born stability is checked post-prediction. No quantity is defined in terms of its own prediction, no parameter is fitted to the test set then 'predicted', and no load-bearing claim rests on a self-citation. The design rules (VEC 6.0–6.4, δ < 10%, Re essential) are empirical summaries of the test-set predictions, which is not circular (though it is an extrapolation concern). The only noted issue is an internal inconsistency: Table 3.3 lists CrMo3 (no Re) as rank 2, while Sec. 3.4 and the abstract claim all top alloys contain Re; this is an arithmetic/consistency error, not a circularity. Self-citations [14,16] concern unrelated prior work (GPR for BCC iron, FeV under pressure) and do not provide premises for the present model. The 100% Born compliance is achieved because all training data are Born-stable and features encode BCC stability descriptors; this is a data property, not a circular argument.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The framework rests entirely on external DFT data and standard elasticity theory; the paper adds a fitted ML model and post-hoc stability checks. The only ad hoc assumption is the representativeness of the 59-alloy test set for global design rules.

free parameters (3)
  • Gradient boosting hyperparameters = n_estimators=700, learning_rate=0.03, max_depth=3, min_samples_split=2, subsample=0.8
    Chosen via 5-fold cross-validation but the search grid and stability are not reported; predictions and Born compliance depend on these values.
  • Train/test split seed = random seed = 7
    The 85/15 stratified split with seed 7 determines the test set; a different seed could change R² and the 100% Born compliance rate.
  • Top-candidate thresholds = top 5 and top 10
    The design rules (Re in 100% of top-10) depend on the arbitrary cutoff; with the top 5 shown in Table 3.3, Re appears in 4 of 5 (80%).
assumptions (5)
  • standard math Born stability criteria for cubic crystals (C11-C12>0, C44>0, C11+2C12>0) are necessary and sufficient for mechanical stability.
    Used in Section 2.6.2 to define stability; standard elasticity result.
  • domain assumption The Zhang et al. [9] DFT dataset (PBE, 16-atom SQS, 0 K) accurately represents the elastic properties of refractory CCAs.
    The model is trained and evaluated entirely on this external dataset; Section 2.2 summarizes validation checks but does not produce new DFT data.
  • domain assumption The 13 compositional descriptors (VEC, delta, elemental statistics) are sufficient to predict the eight mechanical properties.
    Section 2.4 defines the feature set; no structural or electronic-structure descriptors are used, so any composition-structure-property link not encoded in these averages is ignored.
  • domain assumption The polycrystalline isotropy relations E = 9BG/(3B+G) and nu = (3B-2G)/(2(3B+G)) apply to the predicted B and G.
    Used in Section 2.6.1 to derive E and nu from predicted B and G.
  • ad hoc to paper The 59 test alloys are representative of the trillion-scale BCC composition space.
    Required for the screening conclusions in Section 3.5; the paper provides no coverage or sampling analysis to justify generalization.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Physics-Informed Machine Learning for Refractory Alloy Design." pith.science (2026). https://pith.science/paper/3SODRCSD

@misc{pith2026260803805,
  author       = {Pith},
  title        = {Pith review of: Physics-Informed Machine Learning for Refractory Alloy Design},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3SODRCSD}},
  note         = {Machine review of arXiv:2608.03805}
}
read the original abstract

Refractory complex concentrated alloys (CCAs) exhibit exceptional mechanical properties at elevated temperatures, but their vast compositional space (approximately 10^15 possible combinations) poses significant challenges for experimental exploration. We develop a physics-informed machine learning framework combining quantile gradient boosting with Born stability constraints to predict eight mechanical properties (bulk modulus B, shear modulus G, Vickers hardness Hv, elastic constants C11, C12, C44, Young's modulus E, and Poisson's ratio nu) of refractory CCAs. Using 393 density functional theory (DFT)-computed alloys from a 10-element design space (Cr, Hf, Mo, Nb, Re, Ta, Ti, V, W, Zr), we achieve coefficient of determination (R^2) values of 0.89 to 0.97 with mean absolute errors of 0.85 to 14.74 GPa across six properties on a held-out test set (n = 59). Remarkably, all test predictions satisfy the Born stability criteria (100% compliance), demonstrating the efficacy of physics-informed constraints. Compositional screening in valence electron concentration (VEC)-atomic size mismatch (delta) space identifies high-performance candidates, with MoReW exhibiting the highest predicted shear rigidity (C44 = 139 GPa). Elemental analysis reveals that rhenium appears in 100% of the top-performing alloys, molybdenum in 90%, and chromium in 80%, establishing quantitative design rules: VEC = 6.0 to 6.4, delta < 10%, and Re-Mo-Cr ternary systems maximize shear resistance. This physics-informed screening framework enables accelerated discovery of mechanically stable, high-performance refractory alloys from the trillion-scale compositional space.

Figures

Figures reproduced from arXiv: 2608.03805 by the authors.

Figure 3
Figure 3. shows learning curves from [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figure 3.1
Figure 3.1. Learning Curves: Model performance vs. training data size (Note: Small training-validation gap indicates good generalization. Convergence at 250 samples shows data efficiency). achieve sub-GPa precision (MAE = 0.85 GPa), a significant accomplishment given the empirical nature of hardness models. Prediction interval coverage probabilities range from 78.0% to 89.8%, with C12 achieving near-ideal calibration (PICP = 89… view at source ↗
Figure 3.2
Figure 3.2. Predicted versus actual values for six mechanical properties on test set (n = 59). Panels show: (a) bulk modulus B (R2 = 0.974), (b) shear modulus G (R2 = 0.934), (c) Vickers hardness Hv (R2 = 0.892), (d) elastic constant C11 (R2 = 0.960), (e) elastic constant C12 (R2 = 0.930), (f) elastic constant C44 (R2 = 0.888). Solid line: ideal prediction (y = x). Strong agreement across all properties validates the quantile g… view at source ↗
Figures from the paper (2 more)
Figure 3
Figure 3. Figure 3: presents compositional screening maps in VEC [PITH_FULL_IMAGE:figures/full_fig_p014_3.png]
Figure 3.3
Figure 3.3. Figure 3.3: Compositional screening maps in VEC–δ space for 59 test alloys. (a) Born stability: all alloys are mechanically stable (blue markers). (b) C44 magnitude: marker size proportional to shear rigidity. High-C44 zone (VEC = 5.8–6.4, δ = 3–9%) identifies promising candidat…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 28 canonical work pages

  1. [1]

    Senkov, J

    O. Senkov, J. Scott, S. Senkova, D. Miracle, C. Woodward, Microstruc- ture and room temperature properties of a high -entropy tanbhfzrti alloy, Journal of alloys and compounds 509 (20) (2011) 6043–6048

    work page 2011

  2. [2]

    D. B. Miracle, O. N. Senkov, A critical review of high entropy alloys and related concepts, Acta materialia 122 (2017) 448–511

    work page 2017

  3. [3]

    Cantor, I

    B. Cantor, I. T. Chang, P. Knight, A. Vincent, Microstructural de - velopment in equiatomic multicomponent alloys, Materials Science and Engineering: A 375 (2004) 213–218

    work page 2004

  4. [4]

    Yeh, S.-K

    J.-W. Yeh, S.-K. Chen, S.-J. Lin, J.-Y. Gan, T.-S. Chin, T.-T. Shun, C.- H. Tsau, S.-Y. Chang, Nanostructured high-entropy alloys with multiple principal elements: novel alloy design concepts and outcomes, Advanced engineering materials 6 (5) (2004) 299–303

    work page 2004

  5. [5]

    Senkov, G

    O. Senkov, G. Wilks, D. Miracle, C. Chuang, P. Liaw, Refractory high - entropy alloys, Intermetallics 18 (9) (2010) 1758–1765

    work page 2010

  6. [6]

    Miracle, High entropy alloys as a bold step forward in alloy develop- ment, Nature communications 10 (1) (2019) 1805

    D. Miracle, High entropy alloys as a bold step forward in alloy develop- ment, Nature communications 10 (1) (2019) 1805

    work page 2019

  7. [7]

    Hume-Rothery, W

    W. Hume-Rothery, W. Mabbott, Gilbert, K. Channel Evans, The freez - ing points, melting points, and solid solubility limits of the alloys of sliver and copper with the elements of the b sub-groups, Philosophi- cal Transactions of the Royal Society of London. Series A, Containing Papers of a Mathematical or Physical Character 233 (721 -730) (1934) 1–97. 23

    work page 1934

  8. [8]

    Saunders, A

    N. Saunders, A. P. Miodownik, CALPHAD (calculation of phase dia - grams): a comprehensive guide, Vol. 1, Elsevier, 1998

    work page 1998

Show all 32 references
  1. [9]

    Zhang, X

    C. Zhang, X. Gu, U. Bhandari, J. Lei, S. Guo, M. Kourouma, A. Karoui, S. Yang, Machine learning prediction of the mechanical properties of refractory multicomponent alloys based on a dataset of phase and first principles simulation, Frontiers in Metals and Alloys 1 (2022) 1036656

    2022

  2. [10]

    L. Ward, A. Agrawal, A. Choudhary, C. Wolverton, A general -purpose machine learning framework for predicting properties of inorganic ma - terials, npj Computational Materials 2 (1) (2016) 1–7

    2016

  3. [11]

    C. Wen, Y. Zhang, C. Wang, D. Xue, Y. Bai, S. Antonov, L. Dai, T. Lookman, Y. Su, Machine learning assisted design of high entropy alloys with desired property, Acta Materialia 170 (2019) 109–117

    2019

  4. [12]

    Z. Pei, J. Yin, J. A. Hawk, D. E. Alman, M. C. Gao, Machine -learning informed prediction of high -entropy solid solution formation: Beyond the hume-rothery rules, npj Computational Materials 6 (1) (2020) 50

    2020

  5. [13]

    T. Xie, J. C. Grossman, Crystal graph convolutional neural networks for an accurate and interpretable prediction of material properties, Physical review letters 120 (14) (2018) 145301

    2018

  6. [14]

    B. A. Ayirizia, A. De la Rocha, V. I. Arteaga -Muñiz, Y.-H. Tang, W. A. De Jong, J. A. Muñoz San Martín, Gaussian process regression for ma - chine learning on effective crystal graphs of body -centered cubic iron, Solids 6 (4) (2025) 62. doi:10.3390/solids6040062

    2025 doi

  7. [15]

    Lookman, P

    T. Lookman, P. V. Balachandran, D. Xue, R. Yuan, Active learning in materials science with emphasis on adaptive sampling using uncer - tainties for targeted design, npj Computational Materials 5 (1) (2019) 21

    2019

  8. [16]

    Reyes -Pulido, B

    H. Reyes -Pulido, B. KC, R. S. Kumar, R. J. Hemley, J. A. Muñoz, Thermally frustrated phase transition at high pressure in b2 -ordered fev, AIP Advances 14 (7) (2024)

    2024

  9. [17]

    Kresse, J

    G. Kresse, J. Furthmüller, Efficient iterative schemes for ab initio total - energy calculations using a plane-wave basis set, Physical review B 54 (16) (1996) 11169. 24

    1996

  10. [18]

    J. P. Perdew, K. Burke, M. Ernzerhof, Generalized gradient approxima - tion made simple, Physical review letters 77 (18) (1996) 3865

    1996

  11. [19]

    Zunger, S.-H

    A. Zunger, S.-H. Wei, L. G. Ferreira, J. E. Bernard, Special quasirandom structures, Physical review letters 65 (3) (1990) 353

    1990

  12. [20]

    Y. Tian, B. Xu, Z. Zhao, Microscopic theory of hardness and design of novel superhard crystals, International Journal of Refractory Metals and Hard Materials 33 (2012) 93–106

    2012

  13. [21]

    S. Guo, C. Ng, J. Lu, C. Liu, Effect of valence electron concentration on stability of fcc or bcc phase in high entropy alloys, Journal of applied physics 109 (10) (2011)

    2011

  14. [22]

    O. N. Senkov, D. B. Miracle, K. J. Chaput, J.-P. Couzinie, Development and exploration of refractory high entropy alloys —a review, Journal of materials research 33 (19) (2018) 3092–3128

    2018

  15. [23]

    A. Jain, S. P. Ong, G. Hautier, W. Chen, W. D. Richards, S. Dacek, S. Cholia, D. Gunter, D. Skinner, G. Ceder, et al., Commentary: The materials project: A materials genome approach to accelerating materi - als innovation, APL materials 1 (1) (2013)

    2013

  16. [24]

    Meinshausen, G

    N. Meinshausen, G. Ridgeway, Quantile regression forests., Journal of machine learning research 7 (6) (2006)

    2006

  17. [25]

    Pedregosa, G

    F. Pedregosa, G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss, V. Dubourg, et al., Scikit-learn: Machine learning in python, the Journal of machine Learn- ing research 12 (2011) 2825–2830

    2011

  18. [26]

    M. Born, K. Huang, Dynamical theory of crystal lattices, Oxford uni - versity press, 1996

    1996

  19. [27]

    N. L. Okamoto, K. Yuge, K. Tanaka, H. Inui, E. P. George, Atomic displacement in the crmnfeconi high -entropy alloy–a scaling factor to predict solid solution strengthening, AIP Advances 6 (12) (2016)

    2016

  20. [28]

    Pettifor, Theoretical predictions of structure and related properties of intermetallics, Materials science and technology 8 (4) (1992) 345–349

    D. Pettifor, Theoretical predictions of structure and related properties of intermetallics, Materials science and technology 8 (4) (1992) 345–349. 25

    1992

  21. [29]

    G. E. Karniadakis, I. G. Kevrekidis, L. Lu, P. Perdikaris, S. Wang, L. Yang, Physics-informed machine learning, Nature Reviews Physics 3 (6) (2021) 422–440

    2021

  22. [30]

    P. M. Anderson, J. P. Hirth, J. Lothe, Theory of dislocations, Cambridge University Press, 2017

    2017

  23. [31]

    Senkov, S

    O. Senkov, S. Rao, G. Balbus, R. Wheeler, E. Payton, Microstructure and deformation behavior of morew in the temperature range from 25° c to 1500° c, Materialia 27 (2023) 101688

    2023

  24. [32]

    Kaufmann, C

    K. Kaufmann, C. Zhu, A. S. Rosengarten, D. Maryanovsky, T. J. Har - rington, E. Marin, K. S. Vecchio, Crystal symmetry determination in electron diffraction using machine learning, Science 367 (6477) (2020) 564–568. 26 counterwithouttablesection Supplementary Material Table S1...

    2020

Pith tools

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