In description logics, default knowledge is exclusively treated as incidental rules. However, as few concepts are definable using only strict knowledge, imposing strict definitions leads to terminological knowledge bases that mostly contain partially defined concepts. This is a real problem because such concepts can only be inserted as leaves of the terminology. Moreover, instance recognition is biased as these concepts must be explicitly mentioned as properties of these instances. It follows that partially defined concepts are described with necessary but not sufficient conditions. As a solution to these problems, we propose to integrate defaults in concept definitions and we argue that this is essential for our diagnosis application. We introduce a description language AL δ with default(δ) and exception( ) connectives. The cornerstone of our approach is the introduction of a definitional point of view where a default can be part of a concept definition, whereas in the classical inheritance one it is only viewed as a weak implication. We go on to describe a map between the definition of a concept and its inherited properties, and we show that the combination of these definitional and inheritance levels considerably improves the capabilities of classification processes. In particular this allows us to distinguish sure from probable instances and typical from exceptional instances. Finally we provide a specific operation, object refinement, which consists in enlarging object descriptions with exceptions in order to find additional concepts the object is an instance of. This operation is useful for our diagnosis application.

International audienceLudics is a rebuilding of Linear Logic from the sole concept of interaction on objects called designs, that abstract proofs. Works have been done these last years to reconsider the formalization of Natural Language: a dialogue may be viewed as an interaction between such abstractions of proofs. We give a few examples taken from dialogue modeling but also from semantics or speech acts to support this approach

No abstract

We propose in this paper to use Ludics as a unified framework for the analysis of dialogue and the reasoning system. Not only is Ludics a logical theory, but it may also be built by means of concepts of game theory. We first present the main concepts of Ludics. A design is an abstraction and a generalization of the concept of proof. Interaction between designs is equivalent to cut elimination or modus ponens in logical theories. It appears to be a natural means for representing dialogues and also for reasoning. A design is a set of sequences of alternate actions, similar to a move in game theory. We apply Ludics to argumentative dialogues. We discuss how to model the speech acts of argumentative dialogues in terms of dialogue acts. A dialogue act is given by a Ludics action together with the expression that reveals the action in a turn of speech. We show also how arguments may be stored in a commitment state used for reasoning. Finally we revisit an example of juridical dialogue that has been analyzed by Prakken in a different framework.

No abstract