Ebook: Formal Ontologies Meet Industry
Knowledge modeling and the semantic dimension of information plays an increasingly central role in the network economy today. Theoretical research and actual implementations bring up unexpected problems and issues and there is, moreover, an increasing need for solid theoretical foundations for practical applications of ontologies, based on philosophy, linguistics, artificial intelligence and logic. The fifth International workshop Formal Ontology Meets Industry (FOMI 2011), held in Delft, the Netherlands in July 2011, brings together researchers and practitioners involved in this field, without the restrictions of their usual domains of work: business, medicine, engineering, finance, law, biology, geography, electronics, and many more. The nine papers presented here are divided into three sections: Philosophical Foundations of FOMI, Methodological Approaches, and finally, Data Integration. Subjects covered in part one include defining the general notion of technical artifacts in formal ontologies and a philosophy inspired schema to describe applied ontologies. Part two includes a paper on the context and methodological approach for the support of ontology-based annotations and meta data management of clinical trial resources, as well as a presentation of two formal approaches to similarity, the geometric model and feature matching model. In part three, there is a proposal to apply ontologies for Linked Open Data to monitor operational behavior, particularly supply chain risk analysis; this section also includes a description of Gellish, a formal ontological language.
The fifth international workshop “Formal Ontologies Meet Industry” (FOMI 2011), was held in Delft, the Netherlands, on 7 and 8 July 2011. In the tradition of its predecessors, FOMI 2011 is committed to provide an international forum where academic researchers and industrial practitioners meet to analyse and discuss application issues related to methods, theories, tools, and applications based on formal ontologies.
There is today wide agreement that knowledge modelling and the semantic dimension of information plays an increasingly central role in the networked economy: semantic-based applications aim to provide a framework for information and knowledge sharing, reliable information exchange, meaning negotiation, and coordination between distinct organizations or among members of the same organization.
In realising this role, theoretical research driven by the issues that came up by recent work in the more applied domains, and, more often, actual implementations bring up unexpected problems and issues. Moreover there is an increasing need for solid theoretical foundations of practical applications of ontologies, based on philosophy, linguistics, artificial intelligence, and logic. For this reason, FOMI welcomes researchers and practitioners that do research on these topics without restrictions on the domain they deal with: business, medicine, engineering, finance, law, biology, geography, electronics, etc.
As such, one may expect that academic research on formal ontology and applied research on data management in industry lead to active interaction between both groups. This means that the challenge of realising FOMI's goal is primarily one of finding the right ways to attract both groups and organise an event at which both groups want to focus their interaction. Alternatively one may take that research on formal ontology and on data management in industry are rather different efforts carried out by people with different backgrounds and with different tools and goal. Formal ontology may be taken as an academic enterprise in which deliberately distance is taken from getting fast and useful results. The interaction between this research and efforts in industry then depends on one's perspective on how research in academia and industry are related. When one sees the dissemination of academic research to industry as a linear and gradual process in which scientific results are step-by-step transformed into more technological knowledge before they find their way to industrial applications, then the goal of FOMI is indeed challenging. Bringing formal ontologists and representatives of industry together is then like bringing extremes together, ignoring the spectrum of intermediate activities that connect them. FOMI may then offer a platform for formal ontologists and industry to reach out to each other across this spectrum. Yet since both extremes typically are not sharing the same goals and interests, the exchange may lack mutual understanding, commitment or urgency to come to a working agreement. And why should the two extremes meet and try to understand each other? The spectrum of intermediate activities is in place to gradual transform the results in formal ontology into applied ontologies that are more geared towards the interests of industry, so for improving the interaction between formal ontology and industry, it would be better to check the organisation of the separate local links in this spectrum of activities and fix possible gaps that block the interaction between formal ontology and industry.
Yet the idea that the results of academic research are disseminating to industry via a linear and gradual process is a more traditional perspective for which alternatives exist. In models for industrial innovation, representatives of rather different backgrounds collaborate to explore new possibilities by bringing together their different backgrounds. Participants are willing and able to participate in this collaboration, which is not primarily aimed at direct useful results and which does not specifically require that the participants share a common set of concerns, tools, and aims. What is rather needed is a shared willingness to just explore and build up a set of activities. It requires a visionary decision that eventually such collaboration will create all kinds of useful results, yet on the short term it need not do so. The aim of FOMI fits this third perspective. This perspective presupposes that groups are brought together that are in principle different, and that may define all kinds of collaboration that eventually may lead to useful results. It also requires that there is no need for quick results.
This 5th FOMI workshop actually fits all three perspectives to some extent. It brings together researchers with similar backgrounds and who in part already know one another. It also brings together researchers from a whole spectrum of research activities, from work on basic terms in upper ontologies, via on-going development of applied ontologies, to ontologies that are already in use in industry. Possibly the third perspective is less well represented, or, to put it more positively, needs some support to better come out. With the first five workshops FOMI did bring together formal ontologists and representatives of industry. Taking the next step in the interaction means establishing a vision that collaboration will be fruitful in the long term. As such, FOMI 2011 has discussed useful experiences on (1) experienced problems in ontology application, (2) new insights on known problematic issues, (3) new results and observations in ontology implementation, and (4) lessons learned on the best way to apply ontological methodologies to real situations.
The future of FOMI depends what perspective one favours. When formal ontology and industry is taken as akin, FOMI may grow steadily as the community of ontologists in academia and industry is growing. If formal ontologists and industry interact through a spectrum of intermediate activities, it makes sense to broaden the call and aim at bringing in that whole spectrum while simultaneously relaxing the idea that the extremes of this spectrum should directly interact. If the aim is innovation by exploration, it may make sense to let go of the format that participants present their latest work; after five workshops, participants should have a fair understanding of what research is being done, allowing a switch to more explorative activities.
The keynote speakers of FOMI 2011 were Laura Hollink, Delft University of Technology, the Netherlands, Riichiro Mizoguchi, Institute of Scientific and Industrial Research, Osaka University, Japan, and Florian Probst, SAP Research Darmstadt, Germany.
This volume includes the 9 papers presented at the workshop, divided into 3 thematic parts.
The first part on the Philosophical Foundations of FOMI, includes papers by Borgo et al., Garbacz and Trypuz, and Schneider et al.
Borgo et al. present three perspectives on defining the general notion of technical artefacts in formal ontologies. These perspectives are based on the intuitions that technical artefacts are objects that exist by human intervention; and that technical artefacts are to be contrasted to natural entities. The paper further compares and explores similarities and dissimilarities of those perspectives.
Garbacz and Trypuz present a philosophy inspired schema in which to describe applied ontologies of any kind. They posit that metaphilosophy may provide a number of non-trivial insights for engineering metaontology. Compared to the existing metaontological standards this schema reveals the need for a proper specification of the sources of ontological knowledge.
Schneider et al. demonstrate the economic pay-offs of applying sound design principles and theoretical foundations to the development of ontology-driven systems. Using cases from the medical and pharmaceutical domains, they show how rigorous and methodical use of an upper ontology built upon realist principles can lead to a streamed-lined development process of ontologies and ontology-driven systems.
The second part on Methodological Approaches, includes papers by Grenon et al., Keirstead and van Dam, and an invited contribution by Carrara and Morato.
The paper by Grenon et al. presents the context, sketches the methodological approach for the support of ontology-based annotations and metadata management of clinical trial resources. The approach adopted consists in reusing, whenever possible, reference ontologies found in the biomedical domain to support the logical definitions of terminological elements.
Keirstead and van Dam present a study on the conceptualisations of energy systems which are used in computer modelling. Analysis of survey data reveals that researchers find data collection difficult in almost all circumstances which can be facilitated by a common ontology.
The invited contribution by Carrara and Morato present two formal approaches to similarity, the geometrical model and feature matching model, and discuss the prospects of applying such models to the formalisation of technical functions using the family resemblance concept.
The third part focuses on Data Integration applications, and includes contributions by Hofman, Ruijven, and van Renssen.
Hofman proposes to apply ontologies and architectural patterns for Linked Open Data to monitoring operational behaviour. The paper specifically focuses on the implementation of Linked Open Data for supply chain risk analysis.
Ruijven presents an information technology framework based on the use of an ontology that can be used for explicit and consistent specification, design, engineering, production and maintenance of complex facilities. The framework supports the needs of Systems Engineering for unambiguous and explicit communication about such a facility between project participants, companies, disciplines etc.
Finally, van Renssen describes Gellish, a formal ontological language, its implementation in a universal database and message structure, how that language is extended with a capability for the modelling of textual requirements.
We would like to thank the IAOA Steering Committee for their guidance, the Program Committee and the additional reviewers for the insightful reviews, and the Local Organizing Committee for arranging an enjoyable event. We would also like to thank all the researchers who submitted a paper to the workshop. Finally, the workshop would not have been possible without the generous support of our sponsors: Delft University of Technology, Dutch Research School for Information and Knowledge Systems (SIKS), Benelux Association for Artificial Intelligence (BNVKI) and the International Association for Ontology and its Applications (IAOA).
May 2011
Pieter Vermaas and Virginia Dignum
In this paper three perspectives are presented on defining the general notion of technical artifacts in formal ontologies. These perspectives share two intuitions: that technical artifacts are objects that exist by human intervention; and that technical artifacts are to be contrasted to natural entities. Yet the perspectives are different in the way they spell out these intuitions: the relevant human intervention may range from intentional selection to intentional production; and the contrast between technical artifacts and natural entities may be introduced by a constitution relation or by defining properties that set technical artifacts apart. The three perspectives are compared and their similarities and dissimilarities are explored.
The paper defines a schema in which to describe applied ontologies of any kind. In contradistinction to the main trend in engineering metaontology the main ideas that support this schema are inspired by philosophy. Namely, we look at the domain of applied ontologies from the point of view of a certain metaphilosophical tradition. Nevertheless, the final result of our considerations is an engineering artefact, i.e., an OWL ontology. For the sake of validation we employ it to describe a few exemplary applied ontologies.
Domain ontologies that are created without reference to, or not based upon an upper ontology will undermine future ontology development both within and among domains. Applied ontologies based on sound design principles and theoretical foundations provide advantages compared to ontologies neglecting ontological analysis. In this paper we demonstrate the economic pay-offs of applying these principles to the development of ontology-driven systems. We focus on three topics: reuse of pre-existing resources, modularization of ontologies into separable and reusable components, and harmonization of ontologies and knowledge management systems and ex ante harmonization. We show, using cases from the medical and pharmaceutical domains, how rigorous and methodical use of an upper ontology built upon realist principles better enables parametric adaptation to future objects, cross-ontology interoperability, and accommodates objects of both scientific and social importance. More importantly, the application of those principles leads to a streamed-lined development process of ontologies and ontology-driven systems.
This paper presents the context, sketches the methodological approach, and main aspects of the ontologising of the CDISC terminology standard in support of ontology-based annotations and metadata management of clinical trial resources (especially data and records). Such aim is a long term endeavour which involves stages of refinement and also validation. The main purpose of this paper is to put in place the methodological basis for the work of ontological engineering needed. The most challenging task is to carry out the ontologisation of terms in the CDISC terminology. The approach adopted consists in reusing, whenever possible, reference ontologies found in the biomedical domain to support the logical definitions of terminological elements.
This paper presents a study on the conceptualisations of energy systems which are used in computer modelling. A survey was designed and executed, containing questions about the type of energy systems, the scope of the problem, spatial and temporal dimensions of the work as well as software tools, specific conceptualisations and the difficulty of obtaining data. 50 responses were received, with people from a wide range of backgrounds and working on different types of problems. Key results show that almost all respondents use computer models and that the overall majority of them wrote or customised their own tools and models, which are generally complex and take weeks to months for a new user to become familiar with. More than two-thirds of the respondents claimed that it is either difficult or very difficult to collect the right data and a majority is willing to share input data as well as outcomes of the studies. However, only about half of them indicated that they currently use any shared conceptualisations or ontologies in their work. Analysis of the data reveals that researchers find data collection difficult in almost all circumstances and therefore if a common ontology facilitated easier data access, it could be widely applicable. From the results, we can confirm our hypothesis that there is not yet one widely used conceptualisation for energy systems, and it does not seem likely that a one-size-fits-all solution can be developed. Nevertheless the idea of a generic middle-layer ontology for energy systems, taking into account the standards that people already use in their work, could prove useful to link more detailed schemas and future work will explore how this can be achieved to benefit the larger community.
According to [5], the best way to formalize the notion of technical function is to treat such a notion as a family resemblance concept. A family resemblance concept (the term is due to L. Wittgenstein) is characterized and understood in terms of a network of similarity relations that link the various members of the class. Essential to an analysis of the notion of family resemblance is therefore a rigorous definition of the notion of similarity. We present two of the most authoritative formal approaches to similarity, the geometrical model and feature matching model, and we discuss the prospects of applying such models to the case of technical functions.
The main application of ontologies is currently in Linked Open Data. Open, encompassing data published by government organizations that can be re-used for many applications. Data is published in a structured format like XML (eXtensible Markup Language). RDF (Resource Description Framework) is used for specifying links between different data of different resources. In many applications, the semantics has to be derived from the published data. This paper proposes to apply ontologies and architectural patterns for Linked Open Data to monitoring operational behaviour. Not only enterprises require to monitor their behaviour, but also for government authorities gather lots of data related to law enforcement. The paper specifically focuses on the implementation of Linked Open Data for supply chain risk analysis.
The subject of this paper is an information technology framework based on the use of an ontology that can be used for explicit and consistent specification, design, engineering, production and maintenance of complex facilities, e.g. ships and infrastructure. The framework supports the needs of Systems Engineering for unambiguous and explicit communication about such a facility between project participants, companies, disciplines etc. It further facilitates interoperability, increases efficiency and reduces failure costs. To show the complexity of realizing complex facilities nowadays, the first part of the paper gives the basics of a system as a capital facility and the corresponding Systems Engineering process which results in such capital facilities. Subsequently in the second part this is used to explore the need for an ontology-based framework.
Ontologies can be practically applied in industry and governments to solve the current problem of lack of data integration including also the current inability for computers to interpret textual requirements as are present in standards and regulations. This can be realized through the definition and implementation of a formal ontological language that provides the concept definition models (dictionary-taxonomy) and language structure, and in which knowledge, requirements as well as product and process information can be expressed. This paper describes such a formal ontological language, called Gellish (a further development of ISO 10303-221 and ISO 15926), and its implementation in a universal database and message structure and it describes how that language is extended with a capability for the modeling of textual requirements. This enables usage of such requirements models to guide the design of facilities as well as for computer aided verification of designs and deliverables.