Stefano Borgo

Using knowledge as an object: challenges and implications

Knowledge Management Research & Practice (2012) 10, 202–205. doi:10.1057/kmrp.2012.32 Introduction The discipline of knowledge management (KM) has developed around practical purposes that have, nonetheless, important implications in conceptual and theoretical terms. One of these implications is the notion of knowledge that can or must be adopted. The possibility of designing proper methods and tools to facilitate effective knowledge transfer between individuals and organisations, to allow storage of knowledge in some kind of repository and to facilitate retrieval and reuse of knowledge requires a definition of knowledge, which must be, at the same time, conceptually sound and applicable to practical purposes. In many respects, this is a binding condition: it can be argued that all KM solutions imply an explicit or implicit notion of knowledge. However, the elusive nature of knowledge itself, the complexity and multidisciplinary nature of KM and the variety of its application fields have led to the definition and use of a wide range of notions of knowledge and its possible attributes (Holsapple, 2003; Andriessen, 2008). Consequently, there has been an intense discussion on what can be the most appropriate definition of knowledge to employ in the KM field. The discussion on the different types of knowledge, their interactions and the forms in which they can be cast is very animated today. Over the years, this discussion has made clear the complexity of the meanings one usually ascribes to the term ‘knowledge’. Starting from this observation, this Special Issue aims to contribute to the debate in an alternative way. Indeed, we did not ask whether there was the need for a single shared definition of knowledge nor what this definition could be. Considering the substantially practical purpose of KM, reaching a general consensus on a specific notion of knowledge can be less important than selecting one notion of knowledge on a case-by-case basis, provided that this notion is functional to the particular KM solution that one has to develop and apply. In other words, rather than looking for a theoretically sound and ‘universal’ definition of knowledge that satisfies any possible criteria, we argue that it is more suitable and fruitful to adopt a family of notions from which one can select the most appropriate for the application field at stake. Also, this multiple approach becomes feasible when one adopts modern techniques for concept analysis and classification as provided by the area of applied ontology (Staab & Studer, 2009). This position reflected in the Call for Papers for this Special Issue, is compatible with, if not directly adopted by, the approaches proposed in the papers published in it.

Knowledge objects: a formal construct for material, information and role dependences

Information technology has embraced the ontological approach to expand its capacity to deal with knowledge of different kinds. The subsequent combination of formal concerns and philosophical considerations has led to the development of new knowledge systems that rely on foundational distinctions of entity type such as object, process, property and role. This change has attracted attention to the distinction between types of entity in the enterprise. The result is a compartmentalisation of the information, which, although well motivated and technically fruitful, is not always optimal for knowledge management (KM) tasks where one aims at an integrated view of information. This paper focuses on the notion of knowledge object understood as a formal construct for knowledge modelling and KM systems. The approach starts from formal and ontological analysis with an eye to modelling knowledge at large. The paper motivates the introduction of a notion of knowledge object as a new type of entity that emerges from the explicit interaction of material entities, information entities and roles within an enterprise. The main goals of this work are: to discuss the capacity of knowledge objects to tie knowledge and roles in an (enterprise) context; to model aspects of enterprise knowledge that escape standard ontological approaches; and to describe knowledge objects as a conceptual tool that can be integrated within existing formal systems.

Information and organization in public health institutes: an ontology-based modeling of the entities in the reception-analysis-report phases

BackgroundOntologies are widely used both in the life sciences and in the management of public and private companies. Typically, the different offices in an organization develop their own models and related ontologies to capture specific tasks and goals. Although there might be an overall coordination, the use of distinct ontologies can jeopardize the integration of data across the organization since data sharing and reusability are sensitive to modeling choices.ResultsThe paper provides a study of the entities that are typically found at the reception, analysis and report phases in public institutes in the life science domain. Ontological considerations and techniques are introduced and their implementation exemplified by studying the Istituto Zooprofilattico Sperimentale delle Venezie (IZSVe), a public veterinarian institute with different geographical locations and several laboratories. Different modeling issues are discussed like the identification and characterization of the main entities in these phases; the classification of the (types of) data; the clarification of the contexts and the roles of the involved entities. The study is based on a foundational ontology and shows how it can be extended to a comprehensive and coherent framework comprising the different institute’s roles, processes and data. In particular, it shows how to use notions lying at the borderline between ontology and applications, like that of knowledge object. The paper aims to help the modeler to understand the core viewpoint of the organization and to improve data transparency.ConclusionsThe study shows that the entities at play can be analyzed within a single ontological perspective allowing us to isolate a single ontological framework for the whole organization. This facilitates the development of coherent representations of the entities and related data, and fosters the use of integrated software for data management and reasoning across the company.

Towards integration of ontology and text-extracted data for event coreference reasoning

Recently, systems for automatic extraction of semantic information about events from large textual resources have been made available. These tools generate RDF datasets about the events described in the texts, enabling logical reasoning over the extracted information.. Ontological reasoning can be exploited to implement tasks that improve the quality of the extracted information, as, for example in event coreference (i.e., recognizing whether two textual descriptions refer to the same event). Starting from the observation that state of the art tools for event coreference do not exploit ontological information, in this paper, we propose a method to enrich event coreference detection on text-extracted event data by semantic-based rule reasoning.

Data distribution in public veterinary service: health and safety challenges push for context-aware systems

BackgroundToday’s globalised and interconnected world is characterized by intertwined and quickly evolving relationships between animals, humans and their environment and by an escalating number of accessible data for public health. The public veterinary services must exploit new modeling and decision strategies to face these changes. The organization and control of data flows have become crucial to effectively evaluate the evolution and safety concerns of a given situation in the territory. This paper discusses what is needed to develop modern strategies to optimize data distribution to the stakeholders.Main textIf traditionally the system manager and knowledge engineer have been concerned with the increase of speed of data flow and the improvement of data quality, nowadays they need to worry about data overflow as well. To avoid this risk an information system should be capable of selecting the data which need to be shown to the human operator. In this perspective, two aspects need to be distinguished: data classification vs data distribution. Data classification is the problem of organizing data depending on what they refer to and on the way they are obtained; data distribution is the problem of selecting which data is accessible to which stakeholder. Data classification can be established and implemented via ontological analysis and formal logic but we claim that a context-based selection of data should be integrated in the data distribution application. Data distribution should provide these new features:xa0(a) the organization of situation types distinguishing at least ordinary vs extraordinary scenarios (contextualization of scenarios); (b) the possibility to focus on the data that are really important in a given scenario (data contextualization by scenarios);xa0and (c) the classification of which data is relevant to which stakeholder (data contextualization by users).Short conclusionPublic veterinary services, to efficaciously and efficiently manage the information needed for today’s health and safety challenges, should contextualize and filter the continuous and growing flow of data by setting suitable frameworks to classify data, users’ roles and possible situations.