A hybrid science‐guided machine learning approach for modeling chemical processes: A review

Abstract

This study presents a broad perspective of hybrid process modeling combining the scientific knowledge and data analytics in bioprocessing and chemical engineering with a science-guided machine learning (SGML) approach. We divide the approach into two major categories: ML complements science, and science complements ML. We review the literature relating to the hybrid SGML approach, and propose a systematic classification of hybrid SGML models. For applying ML to improve science-based models, we present expositions of direct serial and parallel hybrid modeling and their combinations, inverse modeling, reduced-order modeling, quantifying uncertainty in the process and even discovering governing equations of the process model. For applying scientific principles to improve ML models, we discuss the science-guided design, learning and refinement. For each subcategory, we identify its requirements, strengths, and limitations, together with their published and potential applications. We also present several examples to illustrate different hybrid SGML methodologies for modeling chemical processes.

Redundancy Analysis to Reduce the High-Dimensional Near-Infrared Spectral Information to Improve the Authentication of Olive Oil

“…Among the machine learning classification procedures (i.e., in the case in which the dependent variable is qualitative or non-numeric) from chemical data, , this study applies supervised classification, as the grouping into classes is previously known. First, linear discriminant analysis (LDA) predicts the membership of data to several a priori-defined classes.…”

“…RDA is an alternative to canonical correlation analysis (CCA) presented by authors such as Rao 27 and Van de Wollenberg; 28 more recently, Legendre et al 29 tested the significance of the redundancy axes in RDA. Among the machine learning classification procedures (i.e., in the case in which the dependent variable is qualitative or non-numeric) from chemical data, 30,31 this study applies supervised classification, as the grouping into classes is previously known. First, linear discriminant analysis (LDA) predicts the membership of data to several a priori-defined classes.…”

Parameter Identification Concept for Process Models Combining Systems Theory and Deep Learning

“…Neural ODEs offer a promising approach for hybrid modeling and system identification [12][13][14][15]. Furthermore, the neural network's architecture can be optimized to represent experimental data better.…”

Machine learning alternative to systems biology should not solely depend on data

“…Importantly, the reported achievement of accurately predicting dynamics up to 8 ‘Lyapunov time’ (time required for the divergence in prediction to be considered large by some measure of separation) still translates to under an hour for biomolecular systems (8 × 4.6 min) [ 47 ], which is still not sufficiently long to be useful for systems biology endeavors, even if it can be similarly accomplished. Nonetheless, we believe the boundary of the AI/ML predictions should continue to be ‘pushed’, but a more accurate and reliable approach may benefit from the incorporation of mechanistic elements into the model [ 48 , 49 ].…”