Multi-Connected Ontologies
📝 Abstract
Ontologies have been used for the purpose of bringing system and consistency to subject and knowledge areas. We present a criticism of the present mathematical structure of ontologies and indicate that they are not sufficient in their present form to represent the many different valid expressions of a subject knowledge domain. We propose an alternative structure for ontologies based on a richer multi connected complex network which contains the present ontology structure as a projection. We demonstrate how this new multi connected ontology should be represented as an asymmetric probability matrix.
💡 Analysis
Ontologies have been used for the purpose of bringing system and consistency to subject and knowledge areas. We present a criticism of the present mathematical structure of ontologies and indicate that they are not sufficient in their present form to represent the many different valid expressions of a subject knowledge domain. We propose an alternative structure for ontologies based on a richer multi connected complex network which contains the present ontology structure as a projection. We demonstrate how this new multi connected ontology should be represented as an asymmetric probability matrix.
📄 Content
Multi-Connected Ontologies
Philip Davies Higher Education Bournemouth and Poole College Bournemouth, UK pdavies@bpc.ac.uk
David Newell Software Systems Research Group Bournemouth University Bournemouth, UK dnewell@bournemouth.ac.uk
Abigail Davies St Johns College Oxford University Oxford, UK abigail.davies@sjc.ox.ac.uk
Damla Karagözlü
Software Systems Research Group
Bournemouth University
Bournemouth, UK
damla.karagozlu@gmail.com
Abstract – Ontologies have been used for the purpose of bringing system and consistency to subject and knowledge areas. We present a criticism of the present mathematical structure of ontologies and indicate that they are not sufficient in their present form to represent the many different valid expressions of a subject knowledge domain. We propose an alternative structure for ontologies based on a richer multi connected complex network which contains the present ontology structure as a projection. We demonstrate how this new multi connected ontology should be represented as an asymmetric probability matrix.
Keywords – adaptive, semantic, ontology. I. INTRODUCTION
The present state of ontologies There has been exceptional growth in the annotation of information prompted by the increasing need to share data and study objects based on their structure and semantics. (Gruber 1993) We now find annotated information in a wide range of areas such as language, biology, computing, medicine, web content, etc. Annotated information is created from structured vocabularies known as ontologies. Many disciplines have now developed their own standardized ontologies to enable the sharing of information in their fields. SNOMED, for instance has been produced in the field of medicine, (Price and Spackman 2000) as well as many others which are now being referenced (Noy and McGuinness n.d.).
An ontology defines a common vocabulary for researchers who need to share information in a domain. Many subject areas are now developing ontologies so that specialists can share information in their fields not only with other specialists but even with machines. (Protégé n.d.) Machine-interpretable definitions of basic concepts in the domain and relations among them enable the widespread use of information on the internet and the construction of expert systems.
An ontology uses relationships to organize concepts into hierarchies or subject domains. (Noy and McGuinness n.d.) This paper investigates the present structure of ontologies and whether they are applicable to describing subject domains in their present form. The basic problem we consider is whether the present structure of ontologies is rich enough to represent subject domains fully. We contend that the concept of ontologies needs to be extended in order to fully realise a complete subject domain and we indicate ways in which this extension might be approached
Critique of Ontologies Our approach to ontology structure is drawn from the ideas of the German philosopher Martin Heidegger (1889 – 1976). Heidegger was critical of a one- dimensional division of the world into simplistic categories. According to Heidegger, “The philosophical tradition has misunderstood human experience by imposing a subject-object schema upon it.” (Blatner 2006)
Heidegger gives the example of a hammer which cannot be represented just by its physical features and functions. To understand the hammer you cannot detach it from its relationship to the nails, to the anvil, to the wood, to the experience and skill of the carpenter or to a hundred other things. Just putting it in a category of tools, in an ontology cannot fully capture the human idea of the object and its role in the world. A more complex structure is need to capture the representation of reality. (Blatner 2006)
Robert Pirsig (Pirsig 1974) has also made the point that there always appears more than one workable hypothesis to explain a given phenomenon, and that the number of possible hypotheses appears unlimited. He has developed the idea that there are two types of thinking, the classical and the romantic. The classical way of thinking is characterised by analysing things into their component parts, whereas the romantic sees things as a whole. Classical thought would analyse an object like a motorbike into its physical components; nuts bolts etc. but you can also analyse the motor bike into its functional parts: heat exchanger, generator, exhaust system etc. Pirsig points out that each analysis is equally valid but produces different results. It depends on how you wield the knife of analysis to separate part from part. For example if you take a cylindrical chunk of clay you can cut it straight down and the product is circles, but if you decide to cut at an angle the result is ellipses, if you cut horizontally you obtain rectangles. The result of any analysis is also the product of w
This content is AI-processed based on ArXiv data.