Uncertainty Quantification by Ensemble Learning for Computational Optical Form Measurements
This work addresses uncertainty quantification for computational optical form measurements, which is an incremental improvement by extending an existing deep learning method with ensemble learning.
The paper tackled the problem of quantifying uncertainty in computational optical form measurements by applying ensemble learning to a deep learning approach for solving a large-scale, nonlinear inverse problem, and results showed reliable uncertainty quantification under out-of-distribution errors and noisy data.
Uncertainty quantification by ensemble learning is explored in terms of an application from computational optical form measurements. The application requires to solve a large-scale, nonlinear inverse problem. Ensemble learning is used to extend a recently developed deep learning approach for this application in order to provide an uncertainty quantification of its predicted solution to the inverse problem. By systematically inserting out-of-distribution errors as well as noisy data the reliability of the developed uncertainty quantification is explored. Results are encouraging and the proposed application exemplifies the ability of ensemble methods to make trustworthy predictions on high dimensional data in a real-world application.