Machine Learning in Chemical Product Engineering: The State of the Art and a Guide for Newcomers

Abstract

Chemical Product Engineering (CPE) is marked by numerous challenges, such as the complexity of the properties–structure–ingredients–process relationship of the different products and the necessity to discover and develop constantly and quickly new molecules and materials with tailor-made properties. In recent years, artificial intelligence (AI) and machine learning (ML) methods have gained increasing attention due to their performance in tackling particularly complex problems in various areas, such as computer vision and natural language processing. As such, they present a specific interest in addressing the complex challenges of CPE. This article provides an updated review of the state of the art regarding the implementation of ML techniques in different types of CPE problems with a particular focus on four specific domains, namely the design and discovery of new molecules and materials, the modeling of processes, the prediction of chemical reactions/retrosynthesis and the support for sensorial analysis. This review is further completed by general guidelines for the selection of an appropriate ML technique given the characteristics of each problem and by a critical discussion of several key issues associated with the development of ML modeling approaches. Accordingly, this paper may serve both the experienced researcher in the field as well as the newcomer.

Optimization of Polymer Processing: A Review (Part II-Molding Technologies)

“…This is certainly a trend that still requires further research. For example, Trinh et al [113] applied these types of techniques to chemical product engineering, suggesting some guidelines for further research.…”

The fourth industrial revolution in the food industry—Part I: Industry 4.0 technologies

“…In addition, several studies report that text mining of social media and online communities can be used to automatically identify consumer needs and new product ideas (Kakatkar et al, 2020;Patroni et al, 2020;Zhang et al, 2021). Also, some research has been done on the automatic generation of formulations and process conditions by optimizing predictable quality attributes such as sensory properties, nutrition, and shelf life (Zhang et al 2019;Trinh et al 2021). The latter approach benefits from using hybrid modeling, i.e., a combination of ML and mechanical models.…”

A Comprehensive Study on the Styrene–GTR Radical Graft Polymerization: Combination of an Experimental Approach, on Different Scales, with Machine Learning Modeling

“…Data-driven modeling is based on the premise that patterns and trends, which are otherwise difficult to identify or extract on the basis of knowledge and/or observation, can be identified within data coming from a system or process. Among the most common datadriven approaches that are encountered in the modeling of physicochemical systems are response surface methodology (i.e., as part of design of experiments) and ML methods, the latter being particularly adapted to highly complex or multidimensional problems [35]. In the present work, ten popular supervised ML regression models were initially screened on the produced experimental data of the grafting polymerization system to identify the ones that were more fitted to the specific problem characteristics.…”

“…Machine learning (ML) is a prominent tool in polymer materials design and property prediction that allows saving significant time and effort, related to the development of purely phenomenological modeling approaches, when the underlying phenomena are not completely elucidated and/or highly complex [34,35]. Accordingly, this work presents the implementation of a data-driven modeling approach, on the basis of supervised ML techniques, under varying operating conditions.…”