Abstract. The automatic location of services that fulfill a given need is seen as a key step towards dynamic and scalable integration. In this paper we present a model for the automatic location of services that considers the static and dynamic aspects of service descriptions and identifies what notions of match and techniques are useful for the matching of both. Our model presents three important features: ease of use for the requester, efficient pre-filtering of relevant services, and accurate contracting of services that fulfill a given requester goal. We further elaborate previous work and results on Web service discovery by analyzing what steps and what kind of descriptions are necessary for an efficient and usable automatic service location. Furthermore, we analyze the intuitive and formal notions of match that are of interest for locating services that fulfill a given goal. Although having a formal underpinning, the proposed model does not impose any restrictions on how to implement it for specific applications, but proposes some useful formalisms for providing such implementation. 1

Abstract. Semantic Web Services were developed with the goal of automating the integration of business processes on the Web. The main idea is to express the functionality of the services explicitly, using semantic annotations. Such annotations can, for example, be used for service discovery—the task of locating a service capable of fulfilling a business request. In this paper, we present a framework for annotating Web Services using description logics (DLs), a family of knowledge representation formalisms widely used in the Semantic Web. We show how to realise service discovery by matching semantic service descriptions, applying DL inferencing. Building on our previous work, we identify problems that occur in the matchmaking process due to the open-world assumption when handling incomplete service descriptions. We propose to use autoepistemic extensions to DLs (ADLs) to overcome these problems. ADLs allow for non-monotonic reasoning and for querying DL knowledge bases under local closed-world assumption. We investigate the use of epistemic operators of ADLs in service descriptions, and show how they affect DL inferences in the context of semantic matchmaking. 1

Abstract. The open world assumption makes OWL principally suitable to handle incomplete knowledge in Semantic Web scenarios, however, some scenarios desire closed world reasoning. Autoepistemic description logics allow to realise closed world reasoning in open world settings through epistemic operators. An extension of OWL by epistemic operators therefore allows for non-monotonic features known from closed world systems, such as default rules, integrity constraints or epistemic querying. These features can be beneficially applied in Semantic Web scenarios, where OWL lacks expressiveness. 1

We propose a framework and polynomial algorithms for semantic-based automated Web service composition, fully compliant with Semantic Web technologies. The approach exploits the recently proposed Concept Abduction inference service in Description Logics to extend Concept covering definition to expressive logics and to solve Concept Covering problems in a significant subset of OWL-DL. We show how the proposed approach also deals with not-exact solutions, computing an approximate composition and providing an explanation of which part of the request is not covered by the composite service. We present the formalization of the approach, the proposed algorithms, a prototype system implementing the approach, and illustrate experiments carried out with it. KEY WORDS:

Abstract. The DL-Lite family of tractable description logics lies between the semantic web languages RDFS and OWL Lite. In this paper, we present a probabilistic generalization of the DL-Lite description logics, which is based on Bayesian networks. As an important feature, the new probabilistic description logics allow for flexibly combining terminological and assertional pieces of probabilistic knowledge. We show that the new probabilistic description logics are rich enough to properly extend both the DL-Lite description logics as well as Bayesian networks. We also show that satisfiability checking and query processing in the new probabilistic description logics is reducible to satisfiability checking and query processing in the DL-Lite family. Furthermore, we show that satisfiability checking and answering unions of conjunctive queries in the new logics can be done in LogSpace in the data complexity. For this reason, the new probabilistic description logics are very promising formalisms for data-intensive applications in the Semantic Web involving probabilistic uncertainty. Key words: Bayesian description logics, tractable reasoning, description logics,

Abstract. Semantic service descriptions are frequently given using expressive ontology languages based on description languages. The expressiveness of these languages, however, often implies problems for efficient service discovery, especially when increasing numbers of services become available in large organizations and on the Web. To remedy this problem, we propose an efficient service discovery/retrieval method grounded on a conceptual clustering approach, where services are specified in Description Logics as class definitions [10] and they are retrieved by defining a class expression as a query and by computing the individual subsumption relationship between the query and the available descriptions. We present a new conceptual clustering method that constructs tree indices for clustered services, where available descriptions are the leaf nodes, while inner nodes are intensional descriptions (generalization) of their children nodes. The matchmaking is performed by following the tree branches whose nodes might satisfy the query. The query answering time may strongly improve, since the number of retrieval steps may decrease from O(n) to O(log n) for concise queries. We also show that the proposed method is sound and complete. 1

Semantic service descriptions are frequently given using expressive ontology languages based on description languages. The expressiveness of these languages, however, often implies problems for efficient service discovery, especially when increasing numbers of services become available in large organizations and on the Web. To remedy this problem, we propose an efficient service discovery/retrieval method grounded on a conceptual clustering approach, where services are specified in Description Logics as class definitions [10] and they are retrieved by defining a class expression as a query and by computing the individual subsumption relationship between the query and the available descriptions. We present a new conceptual clustering method that constructs tree indices for clustered services, where available descriptions are the leaf nodes, while inner nodes are intensional descriptions (generalization) of their children nodes. The matchmaking is performed by following the tree branches whose nodes might satisfy the query. The query answering time may strongly improve, since the number of retrieval steps may decrease from O(n) to O(log n) for concise queries. We also show that the proposed method is sound and complete.

Abstract. Although the Internet provides a world wide infrastructure for information provision and communication, the initial web technology stack has substantial draw-backs with regard to automated web content processing. Consequently, the Semantic Web is envisioned as the next evolution step of web technology that shall overcome these deficiencies. Ontologies and Web Services are identified as the key technologies: every resource and every data element interchanged shall be semantically described by ontologies; Web services shall provide access to and usage of computational resources over the Web in order to combine the benefits of the Internet with computational power. The emerging concept of Semantic Web services aims at an integrated technology for amending the Web with semantically enhanced information processing and distributed computing, hence realizing the vision of the Semantic Web. This article explains the idea of Semantic Web services and presents technological solutions developed around the Web Service Modeling Ontology WSMO, a comprehensive framework for Semantic Web Services.

Abstract. Service-oriented computing is gaining wider acceptance. For Web services to become practical, an infrastructure needs to be supported that allows users and applications to discover, deploy, compose and synthesize services automatically. This automation can take place effectively only if formal semantic descriptions of Web services are available. In this paper we present an approach for automatic service discovery and composition with both syntactic and semantic description of Web services. In syntactic case, we use a repository of services described using WSDL (Web Service Description Language). In the semantic case, the services are described using USDL (Universal Service-Semantics Description Language), a language we have developed for formally describing the semantics of Web services. In this paper we show how the challenging task of building service discovery and composition engines can be easily implemented and efficiently solved via (Constraint) Logic programming techniques. We evaluate the algorithms on repositories of different sizes and show the results. 1

Abstract. A framework for modeling Semantic Web Service is proposed. It is based on Description Logic (DL), hence it is endowed with a formal semantics and, in addition, it allows for expressing constraints in service descriptions of different strengths, i.e. Hard and Soft Constraints. Semantic service discovery can be performed by matching DL descriptions, expressing both Hard and Soft constraints, and exploiting DL inferences. Additionally, a method for solving the problem of ranking services is proposed which is based on the use of a semantic similarity measure for DL. This method can rank (matched) service descriptions on the grounds of their semantic similarity w.r.t. the service request, by preferring those that are able to better satisfy both Hard and Soft Constraints.