Semi-supervised Learning Approaches For Microstructure Classification

Transcript

1. 1 Materials Science & Engineering Department Computational Materials Sci. Lab. Semi-supervised Learning Approaches For Microstructure Classification Courtney Kunselman1, Vahid Attari1, Levi McClenny2, Ulisses Braga-Neto2, Raymundo Arroyavea1,3 1Department of Materials Science and Engineering, Texas A&M University 2Department of Electrical Engineering, Texas A&M University 3Department of Mechanical Engineering, Texas A&M University CHiMaD Workshop April 21, 2020
2. 2 Materials Science & Engineering Department Computational Materials Sci. Lab. 2 It is all about exploration! Mars 2020 is a Mars rover mission by NASA's Mars Exploration Program that includes the Perseverance rover with a planned launch on 17 July 2020, and touch down in Jezero crater on Mars on 18 February 2021.
3. 3 Materials Science & Engineering Department Computational Materials Sci. Lab. 3 Why? • 1st element: Acceleration of materials development and deployment • 2nd element: The creation and curation of large scale materials databases has been widely cited as a critical required component for the acceleration of materials development and deployment. – [Niezgoda et al. 2013] • This has been a recognized need by the materials community since the 1970’s . – [Materials Science and Engineering -- Volume II, The Needs, Priorities, and Opportunities for Materials Research] • 3rd element: Forging these links requires quantitative analysis, and while processing parameters and property observations are generally easily quantifiable—they tend to be represented as objects that exist in a relatively low dimensional space – [Kunselman et al. 2019] The need Solution Develop Improve
4. 4 Materials Science & Engineering Department Computational Materials Sci. Lab. 4 The current ongoing work • In particular, we seek to address the following questions: 1. How to we generate microstructure databases? 2. What is the appropriate way to represent the generated high-dimensional microstructure space for inclusion into the database? 3. How do we forge the links between the process-structure-property paradigm in different ensemble of materials and the generated microstructure databases?
5. 5 Materials Science & Engineering Department Computational Materials Sci. Lab. 5 The Strategy for Propagation of Uncertainty Output Input Multi-scale phase-field framework Attari et al Acta Materialia (183), 2020, 452-470
6. 6 Materials Science & Engineering Department Computational Materials Sci. Lab. 6 The Strategy for Propagation of Uncertainty Output Input Multi-scale phase-field framework Attari et al Acta Materialia (183), 2020, 452-470
7. 7 Materials Science & Engineering Department Computational Materials Sci. Lab. 7 Quantification of uncertainty in a CALPHAD model CALPHAD !"#" = % & '()*+ + '-."/0123-2* + '/*24"-3 56 Phase diagram Attari et al Acta Materialia (183), 2020, 452-470 Input parameters '()*+ 7 8-, : = ∑- 8-. 0>- 7 + ?: ∑- 8-@A(8-) + ∑- ∑DEF 8-8D ∑G HI-D 7 (8- − 8D) where νI-D ∅ = νM-D ∅ + νN-D ∅ . :
8. 8 Materials Science & Engineering Department Computational Materials Sci. Lab. 8 Propagation of uncertainty in chain of models: CALPHAD-Microelasticity-Phase-field CALPHAD Microelasticity !"#" = % & '()*+ + '-."/0123-2* + '/*24"-3 56 Phase diagram Elasto-chemical microstructure space Phase-field inputs Attari et al Acta Materialia (183), 2020, 452-470 Composition Kinetic CALPHAD Input parameters 7 88 9: 7 8; 9: 7 << 9: 7 88 9= 7 8; 9= 7 << 9: 6 > 9: 6 > 9= ?-- @
9. 9 Materials Science & Engineering Department Computational Materials Sci. Lab. 9 Propagation of uncertainty in chain of models: Quantities of interests CALPHAD Microelasticity !"#" = % & '()*+ + '-."/0123-2* + '/*24"-3 56 Phase diagram Elasto-chemical microstructure space Phase-field inputs Attari et al Acta Materialia (183), 2020, 452-470 Composition Kinetic CALPHAD Input parameters 7 88 9: 7 8; 9: 7 << 9: 7 88 9= 7 8; 9= 7 << 9: 6 > 9: 6 > 9= ?-- @ Features
10. 10 Materials Science & Engineering Department Computational Materials Sci. Lab. 10 Input space Structure space Characterization with physical descriptors Property space Microstructure quantification and analysis Microelasticity Composition Kinetic CALPHAD c c Mass scattering Interface scattering More anisotropic Finer domain
11. 11 Materials Science & Engineering Department Computational Materials Sci. Lab. 11 General overview of what we did • Forging the process/structure/property links requires quantitative analysis. – This has been a recognized need by the materials community since the 1970’s [Materials Science and Engineering -- Volume II, The Needs, Priorities, and Opportunities for Materials Research]. – The creation and curation of large scale materials databases has been widely cited as a critical required component for the acceleration of materials development and deployment [Niezgoda et al. 2013]. • Pattern recognition and data analysis using the physical descriptors. http://microstructures.net Open Phase-field Microstructure Database
12. 12 Materials Science & Engineering Department Computational Materials Sci. Lab. 12 Motivation • Uncovering links between processing conditions, microstructure, and properties Process Structure Property Forward Propagation Inverse problem Cause and Effect/Trial and Error Goal-Oriented Design Structure space is high-dimensional and difficult to navigate, making its efficient characterization a pressing research interest. Discrete classes sharing structural features are identified and automated classifiers are trained over a feature space which provides adequate discrimination between these classes
13. 13 Materials Science & Engineering Department Computational Materials Sci. Lab. 13 Miscibility boundary engineering Increasing composition for a constant temperature, T Spherical precipitates Bicontinuous Spherical precipitates
14. 14 Materials Science & Engineering Department Computational Materials Sci. Lab. 14 Current status • Supervised learning aims to learn a function that, given a sample of data and desired outputs, approximates a function that maps inputs to outputs. – Need an expert for each specific material system – Expensive (both time and money) – Subjective – Manual labeling • Semi-supervised learning aims to label unlabeled data points using knowledge learned from a small number of labeled data points. • Unsupervised learning does not have (or need) any labeled outputs, so its goal is to infer the natural structure present within a set of data points. • Almost all microstructure classification in the literature comes from: – Well-studied material systems with widely-accepted classes (i.e. steels) – Simulations concerned with building a specific structure (each batch of data is created with a specific class label) • So, there is very little guidance as to how to classify when the ground truth labeling is partially or completely unknown. Classification Supervised Image-level Pixel-level Object-level Semi-supervised Image-level Pixel-level Unsupervised Image-level Pixel-level Segmentation
15. 15 Materials Science & Engineering Department Computational Materials Sci. Lab. 15 Dataset • 10,000 phase field simulations taken from OPMD database Bicontinuous 20% Precipitate 59% Ambiguous 21% DISTRIBUTION OF MICROSTRUCTURE TYPES Bicontinuous Precipitates Ambiguous (Unlabeled)
16. 16 Materials Science & Engineering Department Computational Materials Sci. Lab. 16 Image Processing • Binarization through the Otsu Method – Iterates through all possible thresholds and chooses the one where the sum of background and foreground variances is at a minimum The raw image Binarized version Black autocorrelation function
17. 17 Materials Science & Engineering Department Computational Materials Sci. Lab. 17 Featurization, The microstructure function m(x,l) • For each image, the microstructure function !(#, %) is defined as a wide-sense stationary stochastic process in which ℎ is random variable associated with the probability of finding a specific local state at spatial position #. – ℎ ∈ ), complete set of local states of interest – # is an index, similar to time for a signal • The two-point correlation function is then defined as: *+,+, -. , -/ = 1 ! -. , % ! -/ , %′ , • But stationarity of ! allows us to reduce *+,+3 to a function of one spatial variable 4 = -/ − -. : *+,+, 4 = 1 ! -, % ! - + 4, %′ . • Microstructure function discretizing and periodic boundary conditions and a primitive basis gives: *+,+, 4 = 1 : ; < =< + =<>4 +, ,
18. 18 Materials Science & Engineering Department Computational Materials Sci. Lab. 18 Featurization, The Two-Point Correlation Function, ! ","$ (&) • !","$ (&) is an stochastic microstructure descriptor. • Black phase autocorrelations ((),) (&)) were calculated for each image • This results in a 262,144-dimensional feature space – Dimensionality reduction is necessary: Incremental PCA analysis
19. 19 Materials Science & Engineering Department Computational Materials Sci. Lab. 19 Featurization Normalized Two-Point Correlation Function, !"##$,$&(#) • Why normalized? – Normalizing )*,* # removes the strong relationship with volume fraction, which allows a PCA decomposition of +,-- .,.& # to be used as a discriminative feature space based on structural information. • We define the normalized two-point correlation function as: +,-- .,.& # = 0 1 2, 3 1 2 + #, 35 − 0 1 2, 3 0[1 2 + #, 35 ] 9:-[1 2, 3 ] 9:-[1 2 + #, 35 ] . • For local state < at both endpoints, this reduces to: +,--*,* # = )*,* # − =* > =* − = * > where =* is the black phase volume fraction.
20. 20 Materials Science & Engineering Department Computational Materials Sci. Lab. 20 Dimension Reduction • High-confidence data split into: – Training set (1,536, ~80%) – Validation set (384, ~20%) • IPCA performed on a combination of labeled training and ambiguous sets – Validation set then projected into this space Explained variance of the first five Principal Components (PC) for decompositions of !"," (%) and '(%%b,b(%).
21. 21 Materials Science & Engineering Department Computational Materials Sci. Lab. 21 Supervised Classification, Baseline SVM • Data was mean-centered at zero and scaled to unit variance. • Hyperparameters (! > 0 for all and $ for Gaussian kernel) optimized through exhaustive grid search with 5-fold cross validation on labeled training set. – Soft margin SVM, the larger C gets, the stricter the boundary becomes. • Provides a performance baseline on high-confidence data and a method of visualizing high-dimensional data. ! 10 $ 0.01 Training Error 0.0358 Test Error 0.0547 Collect training data. Assemble features with a property that stores the known class label Instantiate a classifier. Set its parameters if necessary. Train the classifier using the training data. Classify an image or feature collection. Estimate classification error Workflow for classification:
22. 22 Materials Science & Engineering Department Computational Materials Sci. Lab. 22 Thoughts • While the Baseline SVM works well on the high-confidence data, the boundary is uninformed by most of the data points closest to it – Probably cannot be trusted to label the ambiguous set… • We need to assign labels to some of the ambiguous microstructures.
23. 23 Materials Science & Engineering Department Computational Materials Sci. Lab. 23 Semi-Supervised Classification, Transductive SVMs • Uses both labeled and unlabeled data – Unlabeled data, when used in conjunction with a small amount of labeled data, can produce considerable improvement in learning accuracy. • There is always the danger that the wrong semi-supervised method will deteriorate classification performance instead of improving it – To alleviate this concern, a collection of methods with very different mathematical structures are used and the subset identified through their consensus is added to the training set • Supervised method modified to be semi-supervised • Now the problem is to find a boundary in the low density area of the labeled and unlabeled data while still maximizing training accuracy min 1 2 & ' + )* + ,-* . ξ, + )' + 0-* 1 2 ξ0 ∋ 4 50 ∈ −1,1 for < = 1, … , ? • Unfortunately, this problem is NP-hard – But, many methods have been proposed which approximate it
24. 24 Materials Science & Engineering Department Computational Materials Sci. Lab. 24 Semi-Supervised Classifiers • Method 1: Safe Semi-Supervised Support Vector Machine (S4VM), – Transductive SVM approximation algorithm which simultaneously considers multiple low-density separators instead of chasing one local minimum. • Method 2: Label Propagation (LP), – Graph-based method created for the semi-supervised problem • Method 3: COP-KMEANS Clustering (CKM), – Unsupervised method modified to be semi-supervised • Method 4: Modified Yarowsky Algorithm (MY), – Very popular self-training approach because it is easy to understand and can be a wrapper for any existing classifier. • Method 5: Updated training set,
25. 25 Materials Science & Engineering Department Computational Materials Sci. Lab. 25 Semi-supervised Consensus Results
26. 26 Materials Science & Engineering Department Computational Materials Sci. Lab. 26 Updated SVM • The four methods agreed on 301/519 of the initially unlabeled data (~58%) • This subset was added to the initially labeled training set, and an Updated SVM was optimized/trained over this set in the same fashion as the baseline Is there a significant difference? Classifier ! " Training Error (initially labeled) Test Error Baseline 10 0.01 0.0358 0.0547 Updated 10 0.01 0.0397 0.0599
27. 27 Materials Science & Engineering Department Computational Materials Sci. Lab. 27 McNemar’s Test • Non-parametric statistical hypothesis test used to compare dependent categorical outputs – !" : $%& ' = $%) ' • If * + , ≥ 10: χ1 = &2) 3 &%) • If * + , < 10: 5~5789:78;<(* + ,, 0.5) • Per popular convention, looking for significance at the 95% confidence level Conclusion: The difference in performance of baseline and updated SVMs on the high-confidence data is not statistically significant.
28. 28 Materials Science & Engineering Department Computational Materials Sci. Lab. 28 Semi-Supervised Error Estimation • Since there is a set of data for which we do not know the labels, we cannot use traditional error estimation techniques for model validation. A tedious derivation leads to the conclusion: ! = # $ ∈ &' !' + # $ ∈ &) !) ⇒ ̂ ! = , # $ ∈ &' ̂ !' + , # $ ∈ &) ̂ !) . • This means that we can get a semi-supervised error estimate through separate supervised and unsupervised error estimates (with probabilities estimated by total numbers of labeled and labeled sample points). • ̂ !) can be estimated through a traditional test set – but what about ̂ !'? • Labeled subpopulation &) • Unlabeled subpopulation &'
29. 29 Materials Science & Engineering Department Computational Materials Sci. Lab. 29 Unsupervised Error Estimation A recent paper by Platanios et al. outlines a method of unsupervised error estimation through constrained optimization: • Let ! be a set of classifiers and let "# and $# be the agreement rate and error rate of !, respectively – "#: probability that all classifiers in ! assign the same label – $#: probability that all classifiers in ! assign the wrong label • Say ! = {'( , '* , '+ }. Then the solution vector that we are solving for is - = [$ ( , $ * , $ + , $ (,* , $ (,+ , $ *,+ , $ (,*,+ ] E. A. Platanios et al., “Estimating accuracy from unlabeled data,” in Proceedings of the Thirtieth Conference on Uncertainty in Artificial Intelligence, pp. 682– 691, AUAI Press, 2014.
30. 30 Materials Science & Engineering Department Computational Materials Sci. Lab. 30 Unsupervised Error Estimation • Some set theory and combinatorics lead to the conclusion !" = $" + 1 + ' ()* |"| −1 ( ' -⊂" - )( $- , • which gives equality constraints. Inequality constraints include $": " 12 ≤ min 7∈" $"\7 min :∈" $: ≤ 0.5 • Alternatively, a recommended objective function for minimization is: >* ? = ' ": " 12 $" − @ :∈" $: 2 • which minimizes dependence between individual error rates. We also added >2 ? = ' :∈" $: , • which can be optimistic, but takes the dependence of error rates completely out of the problem.
31. 31 Materials Science & Engineering Department Computational Materials Sci. Lab. 31 Semi-supervised Error Estimation • The labeled, unlabeled, and overall error estimates for all five classifiers:
32. 32 Materials Science & Engineering Department Computational Materials Sci. Lab. 32 Summary and Conclusions • Microstructure characterization and classification has been identified as an important step in building processing- structure-property linkages for the ultimate goal of materials by design. • Stochastic descriptors over physical descriptors for microstructure representation. • We – Developed a semi-supervised classification framework which identifies the largest “safe” subset to add to the training set • Data-driven discovery – Defined and utilized a novel semi-supervised error estimation technique • As we move to generalize our work to the more general microstructure classification problem, we must recognize that class taxonomy will be both ambiguous and dynamic. !" # = % &: & () #& − + ,∈& #, ) .! ./ = 0. 20 34/5/67 3! .8 ./ = −9 34/5/67 38
33. 33 Materials Science & Engineering Department Computational Materials Sci. Lab. 33 Thanks http://microstructures.net
34. 34 Materials Science & Engineering Department Computational Materials Sci. Lab.