[14]                                home                        [16]

 

 

 

Key questions on Common Upper Ontology

 

4/20/2004 6:54 AM

also see Objectives of WonderWeb:  http://wonderweb.semanticweb.org/objectives.shtml

 

Note from Leo Obrst,

 

The Role of Foundational Ontologies

WonderWeb: Ontology Infrastructure for the Semantic Web

The WonderWeb Foundational Ontologies Library

 

Irene,

 

There are many issues to address here, but I'll hold off until another time.

 

I'd suggest looking over the WonderWeb Deliverable 18, released last Dec. 2003: Masolo, C., Borgo, S., Gangemi, A., Guarino, N., Oltramari. 2003. A. WonderWeb Deliverable D18 Ontology Library (final), December 31, 2003

 

http://wonderweb.semanticweb.org/deliverables/D18.shtml

 

Here's the first section (quoted from the above), pp. 2-3:

 

1.1 The Role of Foundational Ontologies

 

Ontologies are the basic infrastructure for the Semantic Web. Everybody agrees on this, as the very idea of the Semantic Web hinges on the possibility to use shared vocabularies for describing resource content and capabilities, whose semantics is described in a (reasonably) unambiguous and machine-processable form. Describing this semantics, i.e. what is sometimes called the intended meaning of vocabulary terms, is exactly the job ontologies do for the Semantic Web. But what kinds of ontologies do we need? This is still an open issue. In most practical applications, ontologies appear as simple taxonomic structures of primitive or composite terms together with associated definitions. These are the so-called lightweight ontologies, used to represent semantic relationships among terms in order to facilitate content-based access to the (Web) data produced by a given community. In this case, the intended meaning of primitive terms is more or less known in advance by the members of such community. Hence, in this case, the role of ontologies is more that of supporting terminological services (inferences based on relationships among terms – usually just taxonomic relationships) rather than explaining or defining their intended meaning. On the other hand, however, the need to establishing precise agreements as to the meaning of terms becomes crucial as soon as a community of users evolves, or multicultural and multilingual communities need to exchange data and services. As recently reported by the Harvard Business Review1, this problem may have been “one of the main reasons that so many online market makers have foundered. The transactions they had viewed as simple and routine actually involved many subtle distinctions in terminology and meaning”. To capture (or at least approximate) such subtle distinctions we need an explicit representation of the so-called ontological commitments about the meaning of terms, in order to remove terminological and conceptual ambiguities. A rigorous logical axiomatisation seems to be unavoidable in this case, as it accounts not only for the relationships between terms, but – most importantly – for the formal structure of the domain to be represented. This allows one to use axiomatic ontologies not only to facilitate meaning negotiation among agents, but also to clarify and model the negotiation process itself, and in general the structure of interaction. We should immediately note that building axiomatic ontologies for these purposes may be extremely hard, both conceptually and computationally. However, this job only needs to be undertaken once, before the interaction process starts. The quality of meaning negotiation may drastically affect the trust in a service offered by the Semantic Web, but not the computational performance of the service itself. Thus, for example, a product procurement process involving multiple agents with distributed lightweight ontologies may be carried out in an efficient way by using simple terminological services, but the risk of semantic mismatch can be minimized only if the agents rely on explicit, axiomatised ontologies, which serve to ensure mutual compatibility of the respective models in such a way as to check the extent of real agreement. 1October 2001.

 

1.2: IST Project 2001-33052 WonderWeb: Ontology Infrastructure for the Semantic Web

 

Axiomatic ontologies come in different forms and can have different levels of generality, but a special relevance is enjoyed by the so-called foundational ontologies, which address very general domains. One of the goals of the WONDERWEB project is the development of a library of such foundational ontologies, systematically related to each other in a way that makes the rationales and alternatives underlying different ontological choices as explicit as possible. Hopefully, this library will allow different Semantic Web applications to commit to foundational ontologies according to their own needs and preferences, relying on the chosen modules (and their relationships with the rest of the library) for making explicit the underlying ontological assumptions and their formal consequences. Foundational ontologies are ultimately devoted to facilitate mutual understanding and inter-operability among people and machines. This includes understanding the reasons for non-interoperability, which may in some cases be much more important than implementing an integrated (but unpredictable and conceptually imperfect) system relying on a generic shared “semantics”. In conclusion, we see the role and nature of foundational ontologies (and axiomatic ontologies in general) as complementary to that of lightweight ontologies: the latter can be built semi-automatically, e.g. by exploiting machine learning techniques; the former require more painful human labour, which can gain immense benefit from the results and methodologies of disciplines such as philosophy, linguistics, and cognitive science.

 

1.3 The WonderWeb Foundational Ontologies Library

 

Having motivated the role of foundational ontologies, let us describe the library we have developed within the WONDERWEB project: its philosophy, its structure, and its development approach.