Ontologies in Chemical Engineering: Are we following the right track?

HENNING, GABRIELA P.

Buenos Aires

Congreso; WCCE11 - 11th World Congress on Chemical Engineering; 2023

Asociación Argentina de Ingenieros Químicos

The application of Artificial Intelligence (AI) techniques in Chemical Engineering (CE) has a tradition of more than 5 decades and has gone through different phases [1]. AI is the study of how to make computers do things at which, at the moment, people are better. It encompasses various sub-fields and a variety of techniques. One of the domains that has attracted the interest of Chemical Engineers is the one of explicit knowledge representation, from which ontologies belong to. As it has been pointed out [1], ontologies are the most recent avatar in a long series of knowledge representation concepts, starting with predicate and propositional calculus in the 1940s, semantic networks in the 1960s, frames in the 1970s, and objects in the 1980s.An ontology is an explicit and formal conceptualization that is shared by a community. There have been some recent works associated with the development of CE domain ontologies; the most relevant ones are highlighted in [1]. However, the advances achieved by the CE academic community are still limited, as opposed to what happens in other fields, like bioinformatics and medicine. To the best of the author´s knowledge, with the exception of OntoCAPE [2], which is a comprehensive ontology for the domain of computer-aided process engineering and the pharma-related ontologies, developed by Prof. Ventatasubramanian and coworkers, there are practically no CE ontologies that can be broadly used and which are actually applicable in industry. The development and usage of ontologies is particularly challenging due to variety of reasons, which are pointed out and are thoroughly discussed in this contribution. One of the most significant ones is the lack of adherence to Ontology Engineering (OE) principles. In fact, ontology development processes have improved from the early phases of non-systematic work to current activities performed in the framework of OE. The lack of application of OE principles in the development of many CE ontologies has handicapped the community in making far more substantial contributions in the field. Among others, the most important weaknesses that have been found are linked to the lack of an evaluation stage. Therefore, it is shown how aspects like content, structure, syntax, and semantics can be properly analyzed by adopting suitable methodologies/tools. Finally, an important trend that is taking place in the ontology domain is discussed and presented to the CE community. It refers to some efforts, like the IOF (Industrial Ontology Foundry), which are related to the development of a suite of open and principles-based reference ontologies, from which other domain dependent or application ontologies can be derived in a modular fashion, remaining ‘generic’ (i.e., non-proprietary, non-implementation specific) so they can be reused in various industrial domains.References1.Venkatasubramanian, V. (2019). The promise of artificial intelligence in chemical engineering: Is it here, finally? AIChE Journal, 65(2), 466-478.2.Morbach J., Yang A., Marquardt W. (2007). OntoCAPE—A large-scale ontology for chemical process engineering. Engineering Applications of Artificial Intelligence, 20, 147- 161.