Physics-guided neural network for predicting chemical signatures.

Achieving high classification accuracy on trace chemical residues in active spectroscopic sensing is challenging due to the limited amount of training data available to the classifier. Such classifiers often rely on physics-based models for generating training data though these models are not always accurate when compared to measured data. To overcome this challenge, we developed a physics-guided neural network (PGNN) for predicting chemical reflectance for a set of parameterized inputs that is more accurate than the state-of-the-art physics-based signature model for chemical residues. After training the PGNN, we use it to generate a library of predicted spectra for training a classifier. We compare the classification accuracy when using this PGNN library versus a library generated by the physics-based model. Using the PGNN, the average classification accuracy increases from 0.623 to 0.813 on real chemical reflectance data, including data from chemicals not included in the PGNN training set.