Bøger af Bernhard Ganter

Computerscientistscreatemodelsofaperceivedreality.ThroughAItechniques, these models aim at providing the basic support for emulating cognitive - havior such as reasoning and learning, which is one of the main goals of the AI research e?ort. Such computer models are formed through the interaction of various acquisition and inference mechanisms: perception, concept learning, conceptual clustering, hypothesis testing, probabilistic inference, etc., and are represented using di?erent paradigms tightly linked to the processes that use them. Among these paradigms let us cite: biological models (neural nets, genetic programming), logic-based models (?rst-order logic, modal logic, rule-based s- tems), virtual reality models (object systems, agent systems), probabilistic m- els(Bayesiannets,fuzzylogic),linguisticmodels(conceptualdependencygraphs, language-based representations), etc. OneofthestrengthsoftheConceptualGraph(CG)theoryisitsversatilityin terms of the representation paradigms under which it falls. It can be viewed and therefore used, under di?erent representation paradigms, which makes it a p- ular choice for a wealth of applications. Its full coupling with di?erent cognitive processes lead to the opening of the ?eld toward related research communities such as the Description Logic, Formal Concept Analysis, and Computational Linguistic communities. We now see more and more research results from one community enrich the other, laying the foundations of common philosophical grounds from which a successful synergy can emerge.

This volume contains the proceedings of ICCS 2003, the 11th International C- ferenceonConceptualStructures. Thisconferenceseriescontinuestobethemain forum for the presentation and discussion of state-of-the-art research on conc- tualstructures. Thetheories,methodologies,andtechniquespresentedherehave grown considerably in scope in recent years. On the other hand, solid bridges spanning the boundaries between such diverse ?elds as Conceptual Graphs, F- mal Concept Analysis, and others are increasingly being built in our community. The theme of this year's conference was "e;Conceptual Structures for Kno- edge Creation and Communication"e;. In our increasingly (Inter)networked world, the potential support of information technology for the creation and commu- cation of quality knowledge is almost boundless. However, in reality, many c- ceptual barriers prevent the use of this potential. The main problem is no longer in the technological infrastructure, but in how to navigate, use, and manage the wealth of available data resources. Thus, the question is: how to create and communicate from data the information and ultimately the knowledge required by an ever more complex and dynamic society? Conceptual structures research focuses on what is behind and between the data glut and the information ov- load that need to be overcome in answering this question. In this way, our ?eld contributes important ideas on how to actually realize some of the many still ambitious visions. All regular papers were reviewed in a thorough and open process by at least two reviewers and one editor.

Generalizations that handle incomplete, faulty, orimprecise data are discussed, but the focus lies on knowledge extraction from areliable information source.The method is based on Formal Concept Analysis, a mathematical theory ofconcepts and concept hierarchies, and uses its expressive diagrams.