Hybrid machine learning/physics-based approach for predicting oxide glass-forming ability

Wilkinson, Collin J.; Trivelpiece, Cory L.; Hust, Rob; Welch, Rebecca S.; Feller, Steve; Mauro, John C.

doi:10.1016/j.actamat.2021.117432

Cited by 19 publications

(5 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Linear regression, GPR, and RF were tried for each step, and the RF and GPR models had the lowest error for the two respective steps. Details of these techniques and other uses are described elsewhere 12–14 . The training of the RF was done twice with different features (inputs).…”

Section: Methodsmentioning

confidence: 99%

“…Details of these techniques and other uses are described elsewhere. [12][13][14] The training of the RF was done twice with different features (inputs). The two features used were the absolute temperature (T) and 1000/T, with the 1000/T RF outperforming the T models.…”

Section: Machine Learning Modelsmentioning

confidence: 99%

“…The methods vary in complexity and choosing an appropriate method is determined by the amount of available data one has to train the model, where more data allows for a more complex method. An introduction to ML is available in the work of Friedman et al., 9 and specific applications of ML to glass science are available in Refs 10–13 . The present work provides experimental characterization of the color data of LD glass–ceramics having LT as a secondary phase, and then uses that data as input for a ML algorithm that predicts the properties of the glass–ceramic material from its thermal history.…”

Section: Introductionmentioning

confidence: 99%

“…An introduction to ML is available in the work of Friedman et al, 9 and specific applications of ML to glass science are available in Refs. [10][11][12][13] The present work provides experimental characterization of the color data of LD glassceramics having LT as a secondary phase, and then uses that data as input for a ML algorithm that predicts the properties of the glass-ceramic material from its thermal history. Lastly, a Monte Carlo method was used to solve the inverse problem of determining which thermal history can yield a set of targeted L*a*b* color coordinates.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Experimental analysis and machine learning modeling of optical properties of lithium disilicate glass–ceramics comprising lithium tantalate as a secondary phase

DeCeanne

Wilkinson

Dittmer

et al. 2022

Int J of Appl Glass Sci

Self Cite

View full text Add to dashboard Cite

Color is an important property of lithium disilicate (LD) glass–ceramics for a variety of applications. For example, the color of dental glass–ceramics needs to be tuned to match the color of natural teeth. The present study examines the optical properties of LD glass–ceramics containing lithium tantalate as a secondary phase via ultraviolet–visible spectroscopy. A large data set is analyzed to investigate the relationship between heat treatments and optical properties, specifically the L*a*b* color coordinates. The color coordinates are dependent on the heat treatment following previously acquired phase data. The data are used to train a two‐step machine learning model to predict what heat treatment is needed to obtain a given set of color coordinates.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Machine Learning Modelsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Experimental analysis and machine learning modeling of optical properties of lithium disilicate glass–ceramics comprising lithium tantalate as a secondary phase

DeCeanne

Wilkinson

Dittmer

et al. 2022

Int J of Appl Glass Sci

Self Cite

View full text Add to dashboard Cite

show abstract

“…A further in‐depth discussion of CNNs can be found in numerous sources 25–27,29 . As far as the authors are aware, this is the first application of CNNs to glass‐ceramics; however, we have previously presented other machine learning applications to GC 4,30,31 …”

Section: Introductionmentioning

confidence: 98%

Machine learning predictions of Knoop hardness in lithium disilicate glass‐ceramics

et al. 2023

Self Cite

View full text Add to dashboard Cite

Predicting the effects of ceramic microstructures on macroscopic properties, such as the Knoop hardness, has long been a difficult task. This is particularly true in glass–ceramics, where multiple unique crystalline phases can overlap with a background glassy phase. The combination of crystalline and glassy phases makes it difficult to quantify the percent crystallinity and to predict properties that are the result of the chemical composition and microstructure. To overcome this difficulty and take the first step to build a system for characterizing glass‐ceramics, we predict the Knoop hardness based on scanning electron microscopy images using two computational techniques. The first technique is a computer vision algorithm that allows for physical insights into the system because the features used in a predictive model are extracted from the images. The second technique is machine learning with convolutional neural networks that are trained through transfer learning, allowing for more accurate predictions than the first method but with the downside of being a black box. Discussion of the relative merits of the models is included.

show abstract

Rational design and glass-forming ability prediction of bulk metallic glasses via interpretable machine learning

Long

Pei

2023

J Mater Sci

View full text Add to dashboard Cite

Hybrid machine learning/physics-based approach for predicting oxide glass-forming ability

Cited by 19 publications

References 30 publications

Experimental analysis and machine learning modeling of optical properties of lithium disilicate glass–ceramics comprising lithium tantalate as a secondary phase

Experimental analysis and machine learning modeling of optical properties of lithium disilicate glass–ceramics comprising lithium tantalate as a secondary phase

Machine learning predictions of Knoop hardness in lithium disilicate glass‐ceramics

Rational design and glass-forming ability prediction of bulk metallic glasses via interpretable machine learning

Contact Info

Product

Resources

About