Abstract. We propose a mediator architecture that allows a learning system to retrieve learning objects from heterogeneous repositories. A mediating component accepts queries formulated in a uniform query language, translates them into repository specific queries and passes them to each connected repository. For the translation of queries, a novel ontology-based query-rewriting method has been developed. The architecture has been realized in the Web-based, user-adaptive and interactive e-learning environment ActiveMath. Currently, it enables the ActiveMath's course planner to access four heterogeneous learning object repositories.

Abstract. Semantic Web policies are general statements defining the behavior of a system that acts on behalf of real users. These policies have various applications ranging from dynamic agent control to advanced access control policies. Although policies attracted a lot of research efforts in recent years, suitable representation and reasoning facilities allowing for reactive policies are not likewise developed. In this paper, we describe the concept of reactive Semantic Web policies. Reactive policies allow for the definition of events and actions, that is, they allow to define reactive behavior of a system acting on the Semantic Web. A reactive policy makes use of the tremendous amount of knowledge available on the Semantic Web in order to guide system behaviour while at the same time ensuring trusted and policy-compliant communication. We present a formal framework for expressing and enforcing such reactive policies in combination with advanced trust establishing techniques featuring an interplay between reactivity and agent negotiation. Finally, we explain how our approach was applied in a prototype which allows to define and enforce reactive Semantic Web policies on the Social Network and communication tool Skype.

No abstract

Abstract.Business processes in open distributed environments such as the Web force users to interact with other parties be it users or companies even if they have never had any common transaction in the past. Policy-driven trust negotiations emerged in order to address these situations. But although many policy languages and protocols have been defined, the problem of deciding which credential disclosure set to choose from those that possibly make a negotiation succeed is still subject of research. This paper explores the use of qualitative preferences in order to solve the problem and exploits the recently introduced notions of amalgamated and incremented preferences in order to allow for automated decisions which negotiations are preferred by the user. Our solution eases the task for the user of selection among all possible negotiations by removing irrelevant alternatives and it may even automatize negotiations that otherwise would require manual intervention.

No abstract