Machine learning model training
A method includes receiving spectral data of a substrate and metrology data corresponding to the spectral data of the substrate. The method further includes determining a plurality of feature model configurations for each of a plurality of feature models, each of the plurality of feature model configurations including one or more feature model conditions. The method further includes determining a plurality of feature model combinations, where each feature model combination of the plurality of feature model combinations includes a subset of the plurality of feature model configurations. The method further includes generating a plurality of input datasets, where each input dataset of the plurality of input datasets is generated based on application of the spectral data to a respective feature model combination of the plurality of feature model combinations. The method further includes training a plurality of machine learning models, where each machine learning model is trained to generate an output using an input dataset of the plurality of input datasets and the metrology data. The method further includes selecting a trained machine learning model from the plurality of trained machine learning models satisfying one or more selection criteria.
1. A method comprising:
receiving spectral data of a substrate and metrology data corresponding to the spectral data of the substrate;
determining a plurality of feature model configurations for each of a plurality of feature models, each of the plurality of feature model configurations comprising one or more feature model conditions;
determining a plurality of feature model combinations, wherein each feature model combination of the plurality of feature model combinations comprises a subset of the plurality of feature model configurations;
generating a plurality of input datasets, wherein each input dataset of the plurality of input datasets is generated based on application of the spectral data to a respective feature model combination of the plurality of feature model combinations;
training a plurality of machine learning models, wherein each machine learning model is trained to generate an output using an input dataset of the plurality of input datasets and the metrology data; and
selecting a trained machine learning model from the plurality of trained machine learning models satisfying one or more selection criteria.
2. The method of claim 1 , wherein the output comprises a critical dimension (CD) profile prediction.
3. The method of claim 1 , wherein the plurality of feature models comprises at least one of a principal component analysis (PCA) model, an independent component analysis (ICA) model, or a fast Fourier transform (FFT) model.
4. The method of claim 1 , wherein the plurality of machine learning models comprises at least one of a multi-layer perception (MLP), a gradient boosted tree (GBT), a random forest, a support vector regression (SVR), a neural network, or a recursion algorithm.
5. The method of claim 1 , further comprising processing the metrology data prior to performing the training, wherein the processing comprises at least one of filtering, smoothing, clustering, or quantizing the metrology data.
6. The method of claim 1 , wherein the one or more selection criteria comprise at least one of an accuracy criterion, a processor utilization criterion, a processing speed criterion, or a memory utilization criterion.
7. The method of claim 1 , further comprising:
determining a selection value for each trained machine learning model of the plurality of trained machine learning models, wherein the selection value comprises at least one of a root mean square error (RMSE) value, R-squared (R2) value, or error variance value; and
wherein the one or more selection criteria comprise a selection value criterion.
8. A non-transitory computer-readable storage medium storing instructions which, when executed, cause a processing device to perform operations comprising:
receiving spectral data of a substrate and metrology data corresponding to the spectral data of the substrate;
determining a plurality of feature model configurations for each of a plurality of feature models, each of the plurality of feature model configurations comprising one or more feature model conditions;
determining a plurality of feature model combinations, wherein each feature model combination of the plurality of feature model combinations comprises a subset of the plurality of feature model configurations;
generating a plurality of input datasets, wherein each input dataset of the plurality of input datasets is generated based on application of the spectral data to a respective feature model combination of the plurality of feature model combinations;
training a plurality of machine learning models, wherein each machine learning model is trained to generate an output using an input dataset of the plurality of input datasets and the metrology data; and
selecting a trained machine learning model from the plurality of trained machine learning models satisfying one or more selection criteria.
9. The non-transitory computer-readable storage medium of claim 8 , wherein the output comprises a critical dimension (CD) profile prediction.
10. The non-transitory computer-readable storage medium of claim 8 , wherein the plurality of feature models comprise at least one of a principal component analysis (PCA) model, an independent component analysis (ICA) model, or a fast Fourier transform (FFT) model.
11. The non-transitory computer-readable storage medium of claim 8 , wherein the plurality of machine learning models comprise at least one of a multi-layer perception (MLP), a gradient boosted tree (GBT), a random forest, a support vector regression (SVR), a neural network, or a recursion algorithm.
12. The non-transitory computer-readable storage medium of claim 8 , the operations further comprising processing the metrology data prior to performing the training, wherein the processing comprises at least one of filtering, smoothing, clustering, or quantizing the metrology data.
13. The non-transitory computer-readable storage medium of claim 8 , wherein the one or more selection criteria comprise at least one of an accuracy criterion, a processor utilization criterion, a processing speed criterion, or a memory utilization criterion.
14. The non-transitory computer-readable storage medium of claim 8 , the operations further comprising:
determining a value for each trained machine learning model of the plurality of trained machine learning models, wherein the value comprises at least one of a root mean square error (RMSE) value, R-squared (R2) value, or error variance value; and
wherein the one or more selection criteria comprise a selection value criterion.
15. A system comprising:
a memory; and
a processing device coupled to the memory, the processing device to:
receive spectral data of a substrate and metrology data corresponding to the spectral data of the substrate;
determine a plurality of feature model configurations for each of a plurality of feature models, each of the plurality of feature model configurations comprising one or more feature model conditions;
determine a feature model combination, wherein the feature model combination comprises a subset of the plurality of feature model configurations;
generate an input dataset, wherein the input dataset is generated based on application of the spectral data to the feature model combination; and
train a plurality of machine learning models, wherein each machine learning model is trained to generate an output using the input dataset and the metrology data; and
select a trained machine learning model from the plurality of trained machine learning models satisfying one or more selection criteria.
16. The system of claim 15 , wherein the output comprises a critical dimension (CD) profile prediction.
17. The system of claim 15 , wherein the plurality of feature models comprise at least one of a principal component analysis (PCA) model, an independent component analysis (ICA) model, or a fast Fourier transform (FFT) model.
18. The system of claim 15 , wherein the plurality of machine learning models comprise at least one of a multi-layer perception (MLP), a gradient boosted tree (GBT), a random forest, a support vector regression (SVR), a neural network, or a recursion algorithm.
19. The system of claim 15 , wherein the one or more selection criteria comprise at least one of an accuracy criterion, a processor utilization criterion, a processing speed criterion, or a memory utilization criterion.
20. The system of claim 15 , the processing device further to:
determine a value for each trained machine learning model of the plurality of trained machine learning models, wherein the value comprises at least one of a root mean square error (RMSE) value, R-squared (R2) value, or error variance value; and
wherein the one or more selection criteria comprise a selection value criterion.