Mr. Mobasheri defended his PhD dissertation entitled “A Model for Representing Metascience to Address Scientific Ontologies” on February 15, 2025 (27 Bahman 1403). The abstract of his dissertation is presented below.
Existing scientific ontologies typically represent descriptive scientific knowledge as a collection of entities, relationships, and scientific laws. Each scientific law specifies a relationship among several variables. However, several important aspects of scientific knowledge are often left unrepresented.
The first concerns general considerations about scientific knowledge, including the structural types of scientific laws, the role of scientific theories in the organization of scientific knowledge, and the relationships of variables with the physical world and with one another. The second concerns scientific reasoning as procedural knowledge and the manner in which knowledge derived from relevant scientific laws can be used to answer a specific question. Consequently, the knowledge represented in scientific and domain-specific ontologies cannot readily support automated reasoning for solving problems, identifying gaps in scientific knowledge, or detecting potential inconsistencies.
This dissertation incorporates these previously overlooked aspects into scientific knowledge and proposes a two-level model for representing the resulting body of knowledge. At the lower level, corresponding to conventional scientific ontologies, the concepts and relationships associated with a particular scientific domain are represented. Variables and the relationships among them expressed by scientific laws are also included at this level.
At the upper level, general considerations about scientific knowledge are represented. This level is not itself part of science; rather, it concerns science. In this dissertation, it is referred to as metascience. It encompasses concepts such as the different types of scientific laws, the role of variables in expressing scientific knowledge, and the role of scientific theories. In the proposed metascience model, a scientific theory is not merely a container for laws of nature; it also plays a decisive role in ensuring the correctness of scientific reasoning. The scientific knowledge represented at the lower level is connected to, and conforms with, the model defined at the metascience level.
To demonstrate the usefulness of the proposed metascience representation model, a reasoning system was developed. Given a system, denoted by s, in an ontology of the physical world and a scientific theory, denoted by T, in an ontology of science, the reasoner can use the content of the theory to analyze system s and determine its output as a function of time.
The reasoner can also extract the structure of system s, insofar as it is relevant to theory T, from the large body of information contained in the ontology. During the analysis of system s, it can reveal potential gaps in the scientific knowledge represented in theory T or identify possible inconsistencies within that knowledge.
This reasoner embodies an example of the procedural knowledge that is commonly overlooked in representations of scientific knowledge. An important characteristic of the reasoner is its independence from domain-specific scientific knowledge. In other words, it contains no scientific propositions related to a particular domain for solving problems and relies solely on the representation model prescribed at the metascience level.
The procedures incorporated into the reasoner are therefore reusable and are not restricted to any particular scientific domain or theory. Factoring out these reasoning procedures demonstrates that metascience provides genuine added value. Experiments conducted on problems from the selected case-study domains confirmed the correctness of the reasoner’s procedures.