Large information repositories as digital libraries, online shops, etc.

Abstract. Ranking is an important concept to avoid empty or overfull and unordered result sets. However, such scoring can only express total orders, which restricts its usefulness when several factors influence result relevance. A more flexible way to express relevance is the notion of preferences. Users state which kind of answers they ‘prefer ’ by adding soft constraints to their queries. Current approaches in the Semantic Web offer only limited facilities for specification of scoring and result ordering. There is no common language element to express and formalize ranking and preferences. We present a comprehensive extension of SPARQL which directly supports the expression of preferences. This includes formal syntax and semantics of preference expressions for SPARQL. Additionally, we report our implementation of preference query processing, which is based on the ARQ query engine.

The Bibster system is an application of the use of semantics in Peer-to-Peer systems, which is aimed at researchers that share bibliographic metadata. In this paper we describe the design and implementation of recommender functionality in the Bibster system which allows personalized access to the bibliographic metadata available in the Peer-to-Peer network. These functions are based on a semantic user profile which is created from content and usage information as well as a similarity function. Furthermore, these functions make use of the semantic topology of the Peer-to-Peer system.

Abstract. Sharing structured data in a P2P network is a challenging problem, especially in the absence of a mediated schema. The standard practice of answering a consecutively rewritten query along the propagation path often results in significant loss of information. On the opposite, the use of mediated schemas requires human interaction and global agreement, both during creation and maintenance. In this paper we present GrouPeer, an adaptive, automated approach to both issues in the context of unstructured P2P database overlays. By allowing peers to individually choose which rewritten version of a query to answer and evaluate the received answers, information-rich sources left hidden otherwise are discovered. Gradually, the overlay is restructured as semantically similar peers are clustered together. Experimental results show that our technique produces very accurate answers and builds clusters that are very close to the optimal ones by contacting a very small number of nodes in the overlay. 1

Abstract. Full-text search is a challenging problem in Peer-to-Peer (P2P) systems. Currently two promising directions to solve this problem are (1) distributed indexes like hash-tables (DHTs) and (2) semantic overlay networks (SONs) which can be divided into systems that cluster peers with similar content based on term overlap and systems that map both the content and queries on a shared semantic data structure. In this paper we present the pNear system that combines DHTs with clustering via term overlap and show that we are able to tackle some important disadvantages that hold for the individual approaches. We evaluate our approach via simulations based on a large and realistic data-set that we have constructed for this purpose, and which will be useful for similar experiments by others. 1

Knowledge management solutions relying on central repositories sometimes have not met expectations, since users often create knowledge ad-hoc using their individual vocabulary and using their own individual IT infrastructure (e.g., their laptop). To improve knowledge management for such decentralized and individualized knowledge work, it is necessary to, first, provide a corresponding decentralized IT infrastructure and to, second, deal with specific problems such as security and semantic heterogeneity. In this paper, we describe the technical peerto -peer platform that we have built and summarize some of our experiences applying the platform in case study for coopetitioning organizations in the tourism sector.

In Peer-to-Peer data sharing systems, peers typically play equal roles and have equal capabilities. From the service perspective, peers provide a single service. For this reason, there is no need to explicitly model services. We argue, that for knowledge sharing systems, we need an effective combination of the capabilities of the peers both in terms of the services as well as the content they provide.

Knowledge sharing in communities has attracted much attention in the field of knowledge management in research and practice. In this paper we outline a view where the community lifecycle is supported at di#erent stages. The central component of our framework is the community ontology SWRC+COIN that describes the typical structure of communities. We exemplarily show how communities in the academic domain can be detected automatically by means of analyzing information flow in a bibliographic Peer-to-Peer system and how the instantiated community knowledge base can be exploited to support cooperative work in the communities.

We address the problem of efficiently ranking the best peers w.r.t. a query with multiple, equally weighted predicates -- conjunctive queries -- in shortcut overlay networks. This problem occurs when routing queries in unstructured peer-to-peer networks, such as in peer-to-peer information retrieval applications.

Abstract — Peer selection for query routing is a core task in peer-to-peer networks. Unstructured peer-to-peer systems (like Gnutella) ignore this problem, leading to an abundance of network traffic. Structured peer-to-peer systems (like Chord) enforce a particular, global way of distributing data among the peers in order to solve this problem, but then encounter problems of network volatility and conflicts with the autonomy of the peer data management. In this paper, we propose a new mechanism, INGA, which is based on the observation that query routing in social networks is made possible by locally available knowledge about the expertise of neighbors and a semantics-based peer selection function. We validate INGA by simulation experiments with different data sets. We compare INGA with competing peer selection mechanisms on resulting parameters like recall, message gain or number of messages produced.