Applications of machine learning method in high-performance materials design: a review

Abstract

As a generalized method of mathematical statistics, machine learning (ML) is playing an increasingly significant role in the realm of materials design. More sophisticated methodologies for in-depth understandings and wide applications have been developed from initially simple data relation mappings. The present work first summarizes the basic technical issues of ML and then systematically reviews the main implementation strategies for ML methods in accelerating materials research and development process in recent years, encompassing three primary aspects. Firstly, it is necessary to establish the relationship between the key characteristic parameters and properties in any given materials system for a better prediction and exploration of new materials. Then, the computational algorithms in materials science need to be optimized to replace complex calculations with model-predicted data. Finally, the ML methods are applied to summarize the one-dimensional property data and two-dimensional microstructural images of materials to establish standardized analysis methods. During this process, the domain knowledge in a specific system is of great significance to improving the prediction accuracy and efficiency of ML methods, whether pre-processing experimental or computational databases. The powerful capability of ML methods to handle high-dimensional data will enable researchers to make more effective decisions in materials design. In the future, the relationship between the microstructure and mechanical properties, which is necessary to establish a more effective search engine for alloys with targeted mechanical properties, will be the focus of ML mechanical properties of alloy materials.

Keywords

Machine learning / materials design / domain knowledge / applications of machine learning

References

[1] Zhou ZH.Machine learning. Springier Nature; 2021.

[2] Mandal S,Venugopal S.Artificial neural network modeling to evaluate and predict the deformation behavior of stainless steel type AISI 304L during hot torsion.Appl Soft Comput2009;9:237-44

[3] Guo Z.Modelling the correlation between processing parameters and properties of maraging steels using artificial neural network.Comput Mater Sci2004;29:12-28

[4] Gavard L,MacKay DJC.Bayesian neural network model for austenite formation in steels.Mater Sci Technol1996;12:453-63

[5] Bailer-jones C,Mackay D.Gaussian process modelling of austenite formation in steel.Mater Sci Technol1999;15:287-94

[6] Liu Y,Wang Z.Predicting creep rupture life of Ni-based single crystal superalloys using divide-and-conquer approach based machine learning.Acta Mater2020;195:454-67

[7] Shen C,Huang M,van der Zwaag S.A generic high-throughput microstructure classification and quantification method for regular SEM images of complex steel microstructures combining EBSD labeling and deep learning.J Mater Sci Technol2021;93:191-204

[8] Zhang Z,Chen S.Weld image deep learning-based on-line defects detection using convolutional neural networks for Al alloy in robotic arc welding.J Manuf Process2019;45:208-16

[9] Ma J,Liao X,Huang Y.Tool wear mechanism and prediction in milling TC18 titanium alloy using deep learning.Measurement2021;173:108554

[10] Su Y,Bai Y,Xie J.Progress in materials genome engineering in China.Acta Metall Sin2020;56:1313-23. (in Chinese)

[11] Himanen L,Foster AS.Data-driven materials science: status, challenges, and perspectives.Adv Sci2019;6:1900808 PMCID:PMC6839624

[12] Agrawal A.Perspective: materials informatics and big data: Realization of the “fourth paradigm” of science in materials science.APL Mater2016;4:053208

[13] Raccuglia P,Adler PDF.Machine-learning-assisted materials discovery using failed experiments.Nature2016;533:73-6

[14] Medasani B,Ding H.Predicting defect behavior in B2 intermetallics by merging ab initio modeling and machine learning.npj Comput Mater2016;2:1

[15] Takahashi A,Tanaka I.Conceptual and practical bases for the high accuracy of machine learning interatomic potentials: application to elemental titanium.Phys Rev Materials2017;1:063801

[16] Yuan R,Balachandran PV.Accelerated discovery of large electrostrains in BaTiO₃-based piezoelectrics using active learning.Adv Mater2018;30:1702884

[17] Wang C,Cui Q,Xu W.Tensile property prediction by feature engineering guided machine learning in reduced activation ferritic/martensitic steels.J Nucl Mater2020;529:151823

[18] Niu B,Wang Q.Dual-phase synergetic precipitation in Nb/Ta/Zr co-modified Fe-Cr-Al-Mo alloy.Intermetallics2020;124:106848

[19] Domínguez LA,Todd I.Prediction and validation of quaternary high entropy alloys using statistical approaches.Mater Sci Technol2015;31:1201-6

[20] Wen C,Wang C.Machine learning assisted design of high entropy alloys with desired property.Acta Mater2019;170:109-17

[21] Chang YJ,Lee WJ.Prediction of the composition and hardness of high-entropy alloys by machine learning.JOM2019;71:3433-42

[22] He L,Akebono H.Machine learning-based predictions of fatigue life and fatigue limit for steels.J Mater Sci Technol2021;90:9-19

[23] Liu P,Antonov S.Machine learning assisted design of γ’-strengthened Co-base superalloys with multi-performance optimization.npj Comput Mater2020;6:334

[24] Zhang J,Xiong Y.Design high-entropy carbide ceramics from machine learning.npj Comput Mater2022;8:678

[25] Qiao L,Wan Y,Zhang G.Finite element-based machine learning approach for optimization of process parameters to produce silicon carbide ceramic complex parts.Ceram Int2022;48:17400-11

[26] Xue D,Yuan R.An informatics approach to transformation temperatures of NiTi-based shape memory alloys.Acta Mater2017;125:532-41

[27] Xia K,Liu C.A novel superhard tungsten nitride predicted by machine-learning accelerated crystal structure search.Sci Bull2018;63:817-24

[28] Yu J,Pan S.Machine learning-guided design and development of metallic structural materials.J Mater Inf2021;1:9

[29] Matsugi K,Morinaga M.An electronic approach to alloy design and its application to Ni-based single-crystal superalloys.Mater Sci Eng A1993;172:101-10

[30] Mehjabeen A,Qiu D.Redefining the β-phase stability in Ti-Nb-Zr alloys for alloy design and microstructural prediction.JOM2018;70:2254-9

[31] Bania PJ.Beta titanium alloys and their role in the titanium industry.JOM1994;46:16-9

[32] Hume-Rothery W.The equilibrium and lattice-spacing relations in the system magnesium-cadmium.Proc R Soc Lond A1940;174:471-86

[33] Zhang YM,Evans JRG.Revisiting Hume-Rothery’s Rules with artificial neural networks.Acta Mater2008;56:1094-105

[34] Kattner UR.The calphad method and its role in material and process development.Tecnol Metal Mater Min2016;13:3-15 PMCID:PMC4912057

[35] Tancret F,Menou E.Designing high entropy alloys employing thermodynamics and Gaussian process statistical analysis.Mater Design2017;115:486-97

[36] Wu W.Applying machine learning to accelerate new materials development.Sci Sin Phys Mech Astron2018;48:107001

[37] Wang Y,Rondinelli JM.Symbolic regression in materials science.MRS Commun2019;9:793-805

[38] Cui C,Cao Y.Physical metallurgy guided deep learning for yield strength of hot-rolled steel based on the small labeled dataset.Mater Design2022;223:111269

[39] Jiang L,Zhang H.Physical mechanism interpretation of polycrystalline metals’ yield strength via a data-driven method: a novel Hall-Petch relationship.Acta Mater2022;231:117868

[40] Zhang Y.A strategy to apply machine learning to small datasets in materials science.npj Comput Mater2018;4:81

[41] Xue D,Hogden J,Xue D.Accelerated search for materials with targeted properties by adaptive design.Nat Commun2016;7:BFncomms11241 PMCID:PMC4835535

[42] Dai D,Wei X.Using machine learning and feature engineering to characterize limited material datasets of high-entropy alloys.Comput Mater Sci2020;175:109618

[43] Rickman JM,Harmer MP.Materials informatics for the screening of multi-principal elements and high-entropy alloys.Nat Commun2019;10:10533 PMCID:PMC6565683

[44] Childs CM.Embedding domain knowledge for machine learning of complex material systems.MRS Commun2019;9:806-20

[45] Murdock RJ,Wang AYT.Is domain knowledge necessary for machine learning materials properties?.Integr Mater Manuf Innov2020;9:221-7

[46] Jain A,Hautier G.Commentary: the materials project: a materials genome approach to accelerating materials innovation.APL Mater2013;1:011002

[47] Choudhary K,Reid ACE.The joint automated repository for various integrated simulations (JARVIS) for data-driven materials design.npj Comput Mater2020;6:440

[48] Pedregosa F,Gramfort A.Scikit-learn: machine learning in Python.J Mach Learn Res2011;12:2825-30Available from: https://jmlr.org/papers/v12/pedregosa11a.html. [Last accessed on 25 Sep 2023]

[49] Chen T.Xgboost: a scalable tree boosting system. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York, NY, USA: Association for Computing Machinery; 2016. pp. 785-94.

[50] Abadi M.TensorFlow: learning functions at scale. In: Proceedings of the 21st ACM SIGPLAN International Conference on Functional Programming. New York, NY, USA: Association for Computing Machinery; 2016. p. 1.

[51] Imambi S,Kanagachidambaresan GR.PyTorch. In: Prakash KB, Kanagachidambaresan GR, editors. Programming with TensorFlow. EAI/Springer innovations in communication and computing. Cham: Springer; 2021. pp. 87-104.

[52] Erickson N,Shirkov A. Autogluon-tabular: robust and accurate automl for structured data. arXiv. [Preprint.] Mar 13, 2020 [accessed 2024 Sep 25]. Available from: https://arxiv.org/abs/2003.06505.

[53] Xie T.Crystal graph convolutional neural networks for an accurate and interpretable prediction of material properties.Phys Rev Lett2018;120:145301

[54] Chen C,Zuo Y,Ong SP.Graph networks as a universal machine learning framework for molecules and crystals.Chem Mater2019;31:3564-72

[55] Wang H,Han J.DeePMD-kit: a deep learning package for many-body potential energy representation and molecular dynamics.Comput Phys Commun2018;228:178-84

[56] Wang Z,Cai J,Li J.Vision for energy material design: a roadmap for integrated data-driven modeling.J Energy Chem2022;71:56-62

[57] Naser MZ.Error metrics and performance fitness indicators for artificial intelligence and machine learning in engineering and sciences.Archit Struct Constr2023;3:499-517

[58] George EP,Ritchie RO.High-entropy alloys.Nat Rev Mater2019;4:515-34

[59] Ye Y,Lu J,Yang Y.High-entropy alloy: challenges and prospects.Mater Today2016;19:349-62

[60] Liu Y,Xiao B.Accelerated development of hard high-entropy alloys with data-driven high-throughput experiments.J Mater Inf2022;2:3

[61] Rao Z,Xie R.Machine learning-enabled high-entropy alloy discovery.Science2022;378:78-85

[62] Wang C,Jiang L,Xie J.A property-oriented design strategy for high performance copper alloys via machine learning.npj Comput Mater2019;5:227

[63] Zhang H,Zhu S,Xie J.Machine learning assisted composition effective design for precipitation strengthened copper alloys.Acta Mater2021;215:117118

[64] Ozaki T,Watanabe S.Beta Ti alloys with low young’s modulus.Mater Trans2004;45:2776-9

[65] Pang C,Shi Y,Dong C.Cluster-plus-glue-atom model and universal composition formulas [cluster](glue atom)_x for BCC solid solution alloys.J Alloys Compd2015;652:63-9

[66] Wang Q,Wang Y,Dong C.β-Ti alloys with low young’s moduli interpreted by cluster-plus-glue-atom model.Metall Mater Trans A2013;44:1872-9

[67] Jiang B,Wen D.Effects of Nb and Zr on structural stabilities of Ti-Mo-Sn-based alloys with low modulus.Mater Sci Eng A2017;687:1-7

[68] Yang F,Wang Q.Cluster-formula-embedded machine learning for design of multicomponent β-Ti alloys with low Young’s modulus.npj Comput Mater2020;6:372

[69] Liu X.Giant strains in non-textured (Bi_1/2Na_1/2)TiO₃-based lead-free ceramics.Adv Mater2016;28:574-8

[70] Yan B,Li B.Bi-directional prediction of structural characteristics and effective thermal conductivities of composite fuels through learning from finite element simulation results.Mater Design2020;189:108483

[71] Clark SJ,Pickard CJ.First principles methods using CASTEP.Z Krist Cryst Mater2005;220:567-70

[72] Hafner J.Ab-initio simulations of materials using VASP: density-functional theory and beyond.J Comput Chem2008;29:2044-78

[73] Jinnouchi R,Kresse G.On-the-fly machine learning force field generation: application to melting points.Phys Rev B2019;100:014105

[74] Jinnouchi R,Karsai F,Bokdam M.Phase transitions of hybrid perovskites simulated by machine-learning force fields trained on the fly with bayesian inference.Phys Rev Lett2019;122:225701

[75] Noé F,Müller K.Machine learning for molecular simulation.Annu Rev Phys Chem2020;71:361-90

[76] Zong H,Ding X,Lookman T.Developing an interatomic potential for martensitic phase transformations in zirconium by machine learning.npj Comput Mater2018;4:103

[77] Sun S,Zhang B.Data-driven discovery of formulas by symbolic regression.MRS Bull2019;44:559-64

[78] Wei Q,Deng L,Dong Z.Discovering a formula for the high temperature oxidation behavior of FeCrAlCoNi based high entropy alloys by domain knowledge-guided machine learning.J Mater Sci Technol2023;149:237-46

[79] Melhem HG.Machine learning and its application to civil engineering systems.Civil Eng Syst1996;13:259-79

[80] Yuan J,Li Z,Zhang P.Domain-knowledge-oriented data pre-processing and machine learning of corrosion-resistant γ-U alloys with a small database.Comput Mater Sci2021;194:110472

[81] Zou C,Wang WY.Integrating data mining and machine learning to discover high-strength ductile titanium alloys.Acta Mater2021;202:211-21

[82] Tian F,Gao X,Song H.A structural modeling approach to solid solutions based on the similar atomic environment.J Chem Phys2020;153:034101

[83] Liu G.Concrete compressive strength prediction using an explainable boosting machine model.Case Stud Constr Mater2023;18:e01845

[84] Hu Q.The endless search for better alloys.Science2022;378:26-7

[85] Wang Z,Tao K,Li J.MatGPT: a vane of materials informatics from past, present, to future.Adv Mater2024;36:2306733