Skip to main content
Erschienen in: Integrating Materials and Manufacturing Innovation 3/2017

Open Access 05.07.2017 | Technical Article

High-Dimensional Materials and Process Optimization Using Data-Driven Experimental Design with Well-Calibrated Uncertainty Estimates

verfasst von: Julia Ling, Maxwell Hutchinson, Erin Antono, Sean Paradiso, Bryce Meredig

Erschienen in: Integrating Materials and Manufacturing Innovation | Ausgabe 3/2017

Einloggen, um Zugang zu erhalten

Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.

search-config
loading …

Abstract

The optimization of composition and processing to obtain materials that exhibit desirable characteristics has historically relied on a combination of domain knowledge, trial and error, and luck. We propose a methodology that can accelerate this process by fitting data-driven models to experimental data as it is collected to suggest which experiment should be performed next. This methodology can guide the practitioner to test the most promising candidates earlier and can supplement scientific and engineering intuition with data-driven insights. A key strength of the proposed framework is that it scales to high-dimensional parameter spaces, as are typical in materials discovery applications. Importantly, the data-driven models incorporate uncertainty analysis, so that new experiments are proposed based on a combination of exploring high-uncertainty candidates and exploiting high-performing regions of parameter space. Over four materials science test cases, our methodology led to the optimal candidate being found with three times fewer required measurements than random guessing on average.
Literatur
1.
Zurück zum Zitat Roy R (2010) A primer on the Taguchi method. Soc Manuf Eng, 1–245 Roy R (2010) A primer on the Taguchi method. Soc Manuf Eng, 1–245
2.
Zurück zum Zitat Fisher R A (1921) On the probable error of a coefficient of correlation deduced from a small sample. Metron 1:3–32 Fisher R A (1921) On the probable error of a coefficient of correlation deduced from a small sample. Metron 1:3–32
3.
Zurück zum Zitat Chaloner K, Verdinelli I (1995) Bayesian experimental design: a review. Stat Sci 10(3):273–304CrossRef Chaloner K, Verdinelli I (1995) Bayesian experimental design: a review. Stat Sci 10(3):273–304CrossRef
4.
Zurück zum Zitat Chernoff H (1959) Sequential design of experiments. Ann Math Stat 30(3):755–770CrossRef Chernoff H (1959) Sequential design of experiments. Ann Math Stat 30(3):755–770CrossRef
5.
Zurück zum Zitat Cohn D A, Ghahramani Z, Jordan M I (1996) Active learning with statistical models. J Artif Intell Res 4(1):129–145 Cohn D A, Ghahramani Z, Jordan M I (1996) Active learning with statistical models. J Artif Intell Res 4(1):129–145
6.
Zurück zum Zitat Martinez-Cantin R (2014) BayesOpt: a Bayesian optimization library for nonlinear optimization, experimental design and bandits. J Mach Learn Res 15(1):3735–3739 Martinez-Cantin R (2014) BayesOpt: a Bayesian optimization library for nonlinear optimization, experimental design and bandits. J Mach Learn Res 15(1):3735–3739
7.
Zurück zum Zitat Shan S, Wang GG (2010) Survey of modeling and optimization strategies to solve high-dimensional design problems with computationally-expensive black-box functions. Struct Multidiscip Optim 41(2):219–241. doi:10.1007/s00158-009-0420-2 CrossRef Shan S, Wang GG (2010) Survey of modeling and optimization strategies to solve high-dimensional design problems with computationally-expensive black-box functions. Struct Multidiscip Optim 41(2):219–241. doi:10.​1007/​s00158-009-0420-2 CrossRef
9.
10.
Zurück zum Zitat Ueno T, Rhone T D, Hou Z, Mizoguchi T, Tsuda K (2016) Combo: an efficient bayesian optimization library for materials science. Mater Discov 4:18–21CrossRef Ueno T, Rhone T D, Hou Z, Mizoguchi T, Tsuda K (2016) Combo: an efficient bayesian optimization library for materials science. Mater Discov 4:18–21CrossRef
11.
Zurück zum Zitat Xue D, Xue D, Yuan R, Zhou Y, Balachandran P, Ding X, Sun J, Lookman T (2017) An informatics approach to transformation temperatures of NiTi-based shape memory alloys. Acta Mater 125:532–541CrossRef Xue D, Xue D, Yuan R, Zhou Y, Balachandran P, Ding X, Sun J, Lookman T (2017) An informatics approach to transformation temperatures of NiTi-based shape memory alloys. Acta Mater 125:532–541CrossRef
13.
Zurück zum Zitat Oliynyk A, Antono E, Sparks T, Ghadbeigi L, Gaultois M, Meredig B, Mar A (2016) High-throughput machine-learning-driven synthesis of full-heusler compounds. Chem Mater 28(20):7324–7331CrossRef Oliynyk A, Antono E, Sparks T, Ghadbeigi L, Gaultois M, Meredig B, Mar A (2016) High-throughput machine-learning-driven synthesis of full-heusler compounds. Chem Mater 28(20):7324–7331CrossRef
15.
Zurück zum Zitat Ho T K (1998) The random subspace method for constructing decision forests. IEEE Trans Pattern Anal Mach Intell 20(8): 832–844CrossRef Ho T K (1998) The random subspace method for constructing decision forests. IEEE Trans Pattern Anal Mach Intell 20(8): 832–844CrossRef
16.
Zurück zum Zitat Efron B (2012) Model selection estimation and bootstrap smoothing. Division of Biostatistics, Stanford University Efron B (2012) Model selection estimation and bootstrap smoothing. Division of Biostatistics, Stanford University
19.
Zurück zum Zitat Bocarsly JD, Levin EE, Garcia CA, Schwennicke K, Wilson SD, Seshadri R (2017) A simple computational proxy for screening magnetocaloric compounds. Chem Mater 29(4):1613–1622CrossRef Bocarsly JD, Levin EE, Garcia CA, Schwennicke K, Wilson SD, Seshadri R (2017) A simple computational proxy for screening magnetocaloric compounds. Chem Mater 29(4):1613–1622CrossRef
20.
Zurück zum Zitat Sparks T, Gaultois M, Oliynyk A, Brgoch J, Meredig B (2016) Data mining our way to the next generation of thermoelectrics. Scr Mater 111:10–15CrossRef Sparks T, Gaultois M, Oliynyk A, Brgoch J, Meredig B (2016) Data mining our way to the next generation of thermoelectrics. Scr Mater 111:10–15CrossRef
21.
Zurück zum Zitat Agrawal A, Deshpande P D, Cecen A, Basavarsu G P, Choudhary A N, Kalidindi S R (2014) Exploration of data science techniques to predict fatigue strength of steel from composition and processing parameters. Integr Mater Manuf Innov 3(1):1–19CrossRef Agrawal A, Deshpande P D, Cecen A, Basavarsu G P, Choudhary A N, Kalidindi S R (2014) Exploration of data science techniques to predict fatigue strength of steel from composition and processing parameters. Integr Mater Manuf Innov 3(1):1–19CrossRef
22.
Zurück zum Zitat Ward L, Agrawal A, Choudhary A, Wolverton C (2016) A general-purpose machine learning framework for predicting properties of inorganic materials. arXiv preprint Ward L, Agrawal A, Choudhary A, Wolverton C (2016) A general-purpose machine learning framework for predicting properties of inorganic materials. arXiv preprint
23.
Zurück zum Zitat van der Maaten L, Hinton G (2008) Visualizing data using t-SNE. J Mach Learn Res 9:2579–2605 van der Maaten L, Hinton G (2008) Visualizing data using t-SNE. J Mach Learn Res 9:2579–2605
24.
Zurück zum Zitat O’Mara J, Meredig B, Michel K (2016) Materials data infrastructure: a case study of the citrination platform to examine data import, storage, and access. JOM 68(8):2031–2034CrossRef O’Mara J, Meredig B, Michel K (2016) Materials data infrastructure: a case study of the citrination platform to examine data import, storage, and access. JOM 68(8):2031–2034CrossRef
Metadaten
Titel
High-Dimensional Materials and Process Optimization Using Data-Driven Experimental Design with Well-Calibrated Uncertainty Estimates
verfasst von
Julia Ling
Maxwell Hutchinson
Erin Antono
Sean Paradiso
Bryce Meredig
Publikationsdatum
05.07.2017
Verlag
Springer International Publishing
Erschienen in
Integrating Materials and Manufacturing Innovation / Ausgabe 3/2017
Print ISSN: 2193-9764
Elektronische ISSN: 2193-9772
DOI
https://doi.org/10.1007/s40192-017-0098-z

Weitere Artikel der Ausgabe 3/2017

Integrating Materials and Manufacturing Innovation 3/2017 Zur Ausgabe

    Marktübersichten

    Die im Laufe eines Jahres in der „adhäsion“ veröffentlichten Marktübersichten helfen Anwendern verschiedenster Branchen, sich einen gezielten Überblick über Lieferantenangebote zu verschaffen.