Ebook: Formal Ontologies Meet Industry
This book is a collection of papers addressing the multi-shaped character of knowledge, new studies and applications in the field of ontology and semantic technology. The semantic dimension of information plays an increasingly central role in a networked, knowledge-centred economy and the need to encode information into computer systems has led to a bias towards Knowledge Engineering (KE) solutions as opposed to Knowledge Management (KM). Although the intersection between KE and KM has led to a general improvement of information systems, the overlap between these two fields has tended to eclipse interest in genuine knowledge processes. For this reason, the fourth FOMI workshop was held to coincide with the 10th European Conference of Knowledge Management (ECKM), with the explicit aim of bringing together KR and KM (sub)-communities, providing a platform for discussion of these topics. Subjects covered include: the evolution of ontologies in accordance with the evolution of the domain they are designed for; the study of the notions of functional composition and decomposition; modular ontologies for architecture; ontology infrastructure for electromagnetics; ontology of beliefs; enterprise modelling; the application of ontology-based methodologies and techniques to knowledge management issues; ontologies for standards and domain ontologies for biomedicine. The double focus on methodological and applicative issues represents the main feature of these FOMI 2009 articles. This comprehensive perspective will advance progress towards new frontiers in information systems and knowledge management, where research and development in formal ontology will play a leading role.
From raw data of experience to linguistic forms, from cognitively structured conceptualizations to valuable social assets, the interest in the different nuances of the concept of ‘knowledge’ have characterized natural and social sciences since the dawn of human thought. Across Ages, several theorizations have been put forward and disputes contended around the theme of the multi shaped character of knowledge: in the last years, a new fervour has concerned the production and exchange of knowledge in the business domain. In this context “knowledge contents” have often been flattened to information encoded into computer systems, enabling activities like storage and retrieval through relational database structures, machine-tractability through dedicated computational languages and algorithm-based elaboration. This general phenomenon reveals an underlying conflation between “knowledge assets” possessed-by and transferred-through human resources of firms and “information contents” embedded into industrial information systems. In this picture, Knowledge Management (KM) methodologies and applications have been largely biased by Knowledge Engineering (KE) solutions, stressing the level of formal representation and computability but discarding the centrality of the cognitive construction of knowledge operated by human agents over information flows. Although the intersection between KE and KM has led to a general improvement of information systems in terms of massive data analysis and maintenance, decision-making strategies, information retrieval and exchange, etc., the overlap between these two fields have contributed to eclipse the interest on genuine knowledge processes (production, sharing and transfer of new knowledge) and on how to manage them.
It is nowadays widely agreed that the semantic dimension of information plays an increasingly central role in a networked knowledge-centred economy: semantic-based applications aim to provide a framework for information sharing, reliable information exchange, enabling negotiation and coordination between distinct organizations or among members of the same organization.
As testified by research and industrial projects, ‘Semantic Technologies’ are bound to a life-cycle constituted by acquisition, retrieval, modelling, reuse, publishing and maintenance of knowledge.
In Computer Science, this particular notion of ‘knowledge’ would correspond to T-box (terminological) statements in a knowledge base (i.e. student is a subclass of person). The assertion component (A-box), namely factual knowledge associated to terminology (i.e. John is a person), is not so central in the present context. KQML (Knowledge Query and Manipulation Language) is one of the most well known language protocols in the field, together with FIPA ACL (Foundation for Intelligent Physical Agents – Agent Communication Language)
In Computer Science, this particular notion of ‘knowledge’ would correspond to T-box (terminological) statements in a knowledge base (i.e. student is a subclass of person). The assertion component (A-box), namely factual knowledge associated to terminology (i.e. John is a person), is not so central in the present context.
KQML (Knowledge Query and Manipulation Language) is one of the most well known language protocols in the field, together with FIPA ACL (Foundation for Intelligent Physical Agents – Agent Communication Language)
Historically framed in the philosophical tradition and “imported” into the KE framework in the late 90's, “Formal Ontology” (as a discipline) has recently contributed to broaden the scope of inquiry into the notion of knowledge, exploring interdisciplinary areas like Cognitive Science, Social Sciences, Biology, etc. Nevertheless, the problem of managing such knowledge, especially in the business domain and industrial scenarios, has only been skimmed by Formal Ontology. In one sense, if formalization of the ontological categories/concepts is a task per se and, as such, it fosters knowledge sharing/transfer, it is also true that sharing doesn't come “for free”, and efforts need to be made to 1) understand the role of formal ontologies in the production of new knowledge; 2) turn formal ontologies into effective tools for a new generation of KM software; 3) develop standard formal ontologies for different industrial domains.
These remarks constitute the main reason why we chose to co-locate the fourth edition of the “Formal Ontologies Meet Industry” workshop (FOMI)
Past editions: 2005 (Verona), 2006 (Trento), 2008 (Torino).
Past editions: 2005 (Verona), 2006 (Trento), 2008 (Torino).
In fact the contributions presented under the scope of FOMI show a variety from both the methodological as well as the stylistic viewpoint. Some of these are more theoretically oriented and are especially concerned with genuine ontological analysis. The paper by Hamdani and Gargouri, with the title “Towards an approach for evolving information systems' ontologies” is focused on the evolution of ontologies in accordance with the evolution of the domain they are designed for. They build an approach based on operators of change that is aimed at keeping track of the evolution of the information system ontology design and at the same time keeping the consistency and the coherence with the domain through the whole life cycle of the system.
In some other papers ontological analysis is still the main concern, but it is carried out within the study of a particular domain.
“Parts, Compositions and Decompositions of Functions in Engineering Ontologies” by Vermaas analyzes the relations of functional composition and decomposition and compares them with the part-whole relation in mereology in order to understand whether these can be seen as a sort of part-whole relation specific for functions of technical artefacts. A clearer understanding of these relations is particularly important to improve engineering reasoning in systems that deal with functions, like CAD-CAM systems or engineering knowledge bases.
A work focused on design and in particular on architectural design is the one by Hois, Bhatt and Kutz, “Modular Ontologies for Architectural Design”. Their claim is that, given the heterogeneity of the information of the specific architectural domain (quantitative spatial constraints, qualitative relations, functionally-dependent conceptualizations), a modular ontological approach is the one best suited to the integration of the different but equally important perspectives given by this varied information. Their approach is based on the theory of ε-connections and is aimed at being applied to smart office environments.
Another interesting domain is that of electromagnetics, which the paper by Esposito, Tarricone, Vallone and Zappatore, “Towards an Ontology Infrastructure for Electromagnetics” deals with. An interesting feature of the ontological framework they propose is that it is based on a publicly available top level ontology and the choice has been determined by the will of providing a modular and layered architecture that should enable knowledge sharing and reusability for the electromagnetics scientific community.
“Do you still want to vote for your favourite politician? Ask Ontobella!”, the article by Garbacz, Lechniak, Kulicki and Trypuz presents an interesting approach based on the philosophy of Roman Ingarden to the ontology of beliefs, a domain that, despite its relevance for artificial intelligence, has not so far received the attention it deserves from scholars in the ontology community. The paper also contains an interesting preliminary formalization and an application to an example taken from political debates.
In contrast, some works are concerned with the application of ontology-based methodologies and techniques to a particular knowledge management related issue.
The contribution of Hadj Tayeb and Noureddine with the title “Ontological representation for Algerian enterprise modelling” has the purpose of representing – through an ontological framework – a variety of enterprise modelling techniques, distinguished by a list of criteria identified from a theoretical study and the analysis of real case studies relative to Algerian companies. The ontological framework is represented with Protégé.
Borgo and Pozza's paper, “Disentangling Knowledge Objects” is concerned with building a KM framework based on ontological techniques. In particular, a new notion is introduced in the paper, that of knowledge object, comprising three perspectives: material, informational and organizational, proposed as a key concept for enterprise modelling and KM in general.
In addition, some articles provide a contribution in terms of the application of formal ontologies to specific industry domains.
In “A First-Order Cutting Process Ontology for Sheet Metal Parts”, Grüninger and Delaval build an ontology in first order logic that extends the ontology of the ISO 18629 (Process Specification Language) to support the representation of cutting processes in 2D (two dimensions) to be applied to the domain of manufacturing sheet metal parts.
The work by Grenon and De Francisco, “Ontology-strength Industry Standards” deals with the telecommunication domain under the scope of the European project SUPER, where standards, technology neutral architectures, best practices and guidelines are collected in a framework called NGOSS (New Generation of Operation Support Systems). The paper shows that, through an ontologization of standards, it is not only possible to represent concepts in a shared vocabulary like XML, but it also contributes to the enhancement of standards development, dissemination and operationalization.
Finally, Corsar, Moss, Sleeman and Sim, in their article “Supporting the Development of Medical Ontologies” present an approach to biomedicine based on the construction of small domain ontologies, tailored to some very specific tasks, that are very efficient from the point of view of tractability and inferencing. These ontologies are then integrated with an alignment meta-ontology, enabling interoperability with other standard medical knowledge sources.
As this overview suggests, the double focus on methodological and applicative issues represents the main feature of FOMI 2009 articles, confirming the leitmotiv of the past years' editions: ontologies and ontology-driven methodologies are not simply considered as “closed” systems but as dynamic modules embedded in knowledge technologies. We think that this comprehensive perspective can advance progress towards new frontiers in information systems and knowledge management, where research and development in Formal Ontology plays a leading role.
R. Ferrario and A. Oltramari
The increasing availability of large amounts of data and the growing capability of accessing and processing them, gives us today unprecedented opportunities to advance in many fields, including science, commerce, social relations, government, and business, through information technologies. However, in order to have computing machines supporting this progress, data must be turned into processable knowledge. This “epistemic ascent” cannot be driven by data themselves, as some technologist suggests, but requires hypotheses, theories, and models. Formal ontology is part of the theoretical framework that technology needs in order to get computing systems working with networked data in a consistent way. Formal ontology, however, roots in still open philosophical hypotheses. Yet, only formal ontology has the key of notions such as parthood or dependence, that are relevant when discovering knowledge into data. This paper wants to argue that, regardless of philosophy, formal ontology can already provide computer science with many useful tools. On the other hand, without stepping into metaphysics, business communities are on the way of using commonsense notions with more formal consciousness. Making more clear the role of formal ontology can help this process and boost progress towards new frontiers in information management.
Information system's ontology is not frozen: it must evolve in the course of its life cycle to keep its consistency and its coherence in dynamic and multi-disciplinary domains. However, any change to be taken care at the level of Information System (IS) design ontology is a wonderful source of new structural and semantic disconnectedness, which must be perfected to keep the ontology operational and consistent.
In this paper, we define an approach for the recognition of the IS design ontology evolution. Change operators are defined for this goal. Besides, strategies of evolution are defined to manage the consequences of these changes.
The semantic integration of manufacturing systems has been impeded by the lack of rigorous ontologies for specific domains of manufacturing processes and resources. In this paper we present a cutting process ontology for 2D shapes such as sheet metal parts, axiomatized in first-order logic. This ontology is an extension of the ontology of ISO 18629 (Process Specification Language) and an earlier shape ontology first used to support object recognition. The full ontology consists of an axiomatization of all possible ways to change a surface as the result of a cutting process and a taxonomy of cutting processes. All component ontologies are verified using representation theorems.
In this paper I explore the possibility of introducing in engineering ontologies a generic part-whole relation for functions of technical artefacts by functional composition or functional decomposition. I show by means of the postulates of mereology that general functional compositions and decompositions cannot define such a relation. Yet, one can argue that functional decompositions that are acceptable in engineering may define a proper generic part-whole relation for functions. This possibility requires that (i) the part-whole relation is relative to the specific organisation of functions in decompositions, (ii) there is no strict symmetry between functional composition and decomposition, and (iii) functional decomposition is not transitive.
Ontologies are more and more adopted to provide knowledge sharing and reuse, and to promote cooperation between several scientific fields. An increasing need of cooperation and resource sharing is emerging in the electromagnetic arena as well. The range of potential semantic-based applications seems to be wide and the knowledge to be codified is complex and diverse. Therefore the implementation of a well structured ontology in the electromagnetic domain is important. This paper describes a proposal for an electromagnetic ontology framework. The main design criteria were reusability and shareability, in order to make electromagnetic knowledge embeddable in larger and more general frameworks. The implementation of an ontology for aperture-antenna array design is provided as a specific example of a practical application.
The Algerian enterprises are faced to the international industrial and economic competition which requires a complete reorganization of their structure. The modeling concept can satisfy this need and led to many techniques. However, none of them is complete and allows modeling all the aspects of a system.
From a theoretical study and real cases of Algerian enterprises, we developed a knowledge framework through a list of criteria related to characteristics of enterprise modeling. In this context, this article deals with a knowledge representation by a domain ontology realized with the Protégé tool.
Designs of architectural environments have to take into account various sources of heterogeneous information. Not only quantitative spatial constraints and qualitative relations but also functionally-dependent and abstract conceptualizations are relevant aspects for an architectural design. We aim at a modular ontological approach based on the theory of E-connections to formally present and bring together these different perspectives on the domain. Modularity here allows a flexible integration of the various sources while keeping their thematically different aspects apart. We show how modular ontologies reflect the domain for architectural design and how they can be applied.
This paper discusses an ontologisation of a subset of the Telemanagement Forum's New Generation Operational Support Systems standards. The result is a set of ontologies, each corresponding to a particular element in the set of standards, covering complementary aspects of the telecommunication domain. These ontologies are articulated in a modular way according to the standards they target. The ontologies are also mapped, again, in ways adapted from the correspondences put forward within the set of standards itself. We show that ontologisation of informally or semi-formally laid standards can involve non-trivial ontological engineering choices. In addition to sharing lessons learnt, our aim is to support the view that the production of ontologies issued from standards is an endeavour that has the potential of furthering and enhancing standards' development, their dissemination, and operationalisation. Ontologisation of standards, we believe, should be part of the standard development lifecycle.
We discuss a framework for knowledge management based on ontological techniques. The central notion we study, called knowledge object, is provided as a first attempt to provide a new approach to model the relationship between the material, the information and the organization's perspectives. The paper comprises two parts. In the first part it gives an introduction to knowledge objects, their elements and some basic properties. In the second part, it shows how knowledge objects help in understanding and clarifying knowledge creation and evolution in the scenario of a veterinary public health institute, namely the Istituto Zooprofilattico Sperimentale delle Venezie (IZSVE).
The paper presents the preliminary version of Ontobella – a domain ontology of beliefs. The philosophical assumptions of this system are taken from the philosophy of Roman Ingarden and from the psychological results obtained in the Lvov-Warsaw school. Ontobella is applied as the conceptual framework for a computer system that collects information about political debates. We use it to retrieve and store beliefs expressed during sessions of the Polish parliament.
Ontologies are widely used in the biomedical community, which has built standard reference ontologies for various aspects of medicine. These projects have produced broad descriptions of the medical domain, resulting in large, complex ontologies which can be difficult to reuse as part of a focused application. We describe four ontologies which we developed to support intelligent reasoning about a particular medical sub-domain. We also describe how concepts in these ontologies can be aligned with standard reference ontologies to promote interoperability.