Recommender systems face up to current information overload by selecting automatically items that match the personal preferences of each user. The so-called content-based recommenders suggest items similar to those the user liked in the past, by resorting to syntactic matching mechanisms. The rigid nature of such mechanisms leads to recommend only items that bear a strong resemblance to those the user already knows. In this paper, we propose a novel content-based strategy that diversifies the offered recommendations by employing reasoning mechanisms borrowed from the Semantic Web. These mechanisms discover extra knowledge about the user's preferences, thus favoring more accurate and flexible personalization processes. Our approach is generic enough to be used in a wide variety of personalization applications and services, in diverse domains and recommender systems. The proposed reasoning-based strategy has been empirically evaluated with a set of real users. The obtained results evidence computational feasibility and significant increases of recommendation accuracy in relation to existing approaches where our reasoning capabilities are disregarded.

@inproceedings{blanco-fernandez2008semantic, author = {Blanco-Fernandez, Yolanda and Pazos-Arias, José J. and Gil-Solla, Alberto and Ramos-Cabrer, Manuel and Lopez-Nores, Martin}, title = {Semantic Reasoning: A Path To New Possibilities of Personalization}, editor = {Hauswirth, Manfred and Koubarakis, Manolis and Bechhofer, Sean}, booktitle = {Proceedings of the 5th European Semantic Web Conference}, series = {LNCS}, publisher = {Springer Verlag}, address = {Berlin, Heidelberg}, year = {2008}, url = {http://data.semanticweb.org/conference/eswc/2008/papers/11}, keywords = {techniques activation associations web personalization semantic reasoning spreading user-interfaces-and-personalization}, abstract = {Recommender systems face up to current information overload by selecting automatically items that match the personal preferences of each user. The so-called content-based recommenders suggest items similar to those the user liked in the past, by resorting to syntactic matching mechanisms. The rigid nature of such mechanisms leads to recommend only items that bear a strong resemblance to those the user already knows. In this paper, we propose a novel content-based strategy that diversifies the offered recommendations by employing reasoning mechanisms borrowed from the Semantic Web. These mechanisms discover extra knowledge about the user's preferences, thus favoring more accurate and flexible personalization processes. Our approach is generic enough to be used in a wide variety of personalization applications and services, in diverse domains and recommender systems. The proposed reasoning-based strategy has been empirically evaluated with a set of real users. The obtained results evidence computational feasibility and significant increases of recommendation accuracy in relation to existing approaches where our reasoning capabilities are disregarded.} }

%0 = inproceedings %A = Blanco-Fernandez, Yolanda and Pazos-Arias, José J. and Gil-Solla, Alberto and Ramos-Cabrer, Manuel and Lopez-Nores, Martin %B = Proceedings of the 5th European Semantic Web Conference %C = Berlin, Heidelberg %D = 2008 %I = Springer Verlag %T = Semantic Reasoning: A Path To New Possibilities of Personalization %U = http://data.semanticweb.org/conference/eswc/2008/papers/11

In the semantic web environment, where two or more independent ontologies can be used in order to describe knowledge and data, ontologies have to be aligned by defining mappings among the elements of one ontology and the elements of another ontology. Very often, mappings are not derived by the semantics of the ontologies that are compared, but, rather, by an evaluation of the similarity of the terminology used in the two ontologies or of their syntactic structure. Moreover, ontology mappings can be inaccurate, because ontology matching tools derive such mappings from inaccurate terminology or even because they are not specifically tailored for the domain at hand. In this paper, we propose a new mapping validation approach for interpreting similarity-based mappings as semantic relations, by coping also with inaccuracy situations. The idea is to see two independent ontologies as a unique distributed knowledge base and to assume a semantic interpretation of ontology mappings as probabilistic and hypothetical relations among ontology elements. We present and use a probabilistic reasoning tool in order to validate mappings and to possibly infer new relations among the ontologies.

@inproceedings{castano2008mapping, author = {Castano, Silvana and Ferrara, Alfio and Lorusso, Davide and Näth, Tobias Henrik and Moeller, Ralf}, title = {Mapping Validation by Probabilistic Reasoning}, editor = {Hauswirth, Manfred and Koubarakis, Manolis and Bechhofer, Sean}, booktitle = {Proceedings of the 5th European Semantic Web Conference}, series = {LNCS}, publisher = {Springer Verlag}, address = {Berlin, Heidelberg}, year = {2008}, url = {http://data.semanticweb.org/conference/eswc/2008/papers/90}, keywords = {matching mapping probabilistic logics description ontology reasoning formal-languages-1}, abstract = {In the semantic web environment, where two or more independent ontologies can be used in order to describe knowledge and data, ontologies have to be aligned by defining mappings among the elements of one ontology and the elements of another ontology. Very often, mappings are not derived by the semantics of the ontologies that are compared, but, rather, by an evaluation of the similarity of the terminology used in the two ontologies or of their syntactic structure. Moreover, ontology mappings can be inaccurate, because ontology matching tools derive such mappings from inaccurate terminology or even because they are not specifically tailored for the domain at hand. In this paper, we propose a new mapping validation approach for interpreting similarity-based mappings as semantic relations, by coping also with inaccuracy situations. The idea is to see two independent ontologies as a unique distributed knowledge base and to assume a semantic interpretation of ontology mappings as probabilistic and hypothetical relations among ontology elements. We present and use a probabilistic reasoning tool in order to validate mappings and to possibly infer new relations among the ontologies.} }

%0 = inproceedings %A = Castano, Silvana and Ferrara, Alfio and Lorusso, Davide and Näth, Tobias Henrik and Moeller, Ralf %B = Proceedings of the 5th European Semantic Web Conference %C = Berlin, Heidelberg %D = 2008 %I = Springer Verlag %T = Mapping Validation by Probabilistic Reasoning %U = http://data.semanticweb.org/conference/eswc/2008/papers/90

The semantics of OWL-DL and its subclasses are based on the classical semantics of first-order logic, in which the interpretation domain may be an infinite set. This constitutes a serious expressive limitation for such ontology languages, since, in many real application scenarios for the Semantic Web, the domain of interest is actually finite, although the exact cardinality of the domain is unknown. Hence, in these cases the formal semantics of the OWL-DL ontology does not coincide with its intended semantics. In this paper we start filling this gap, by considering the subclasses of OWL-DL which correspond to the logics of the DL-Lite family, and studying reasoning over finite models in such logics. In particular, we mainly consider two reasoning problems: deciding satisfiability of an ontology, and answering unions of conjunctive queries (UCQs) over an ontology. We first consider the description logic DL-Lite_R and show that, for the two above mentioned problems, finite model reasoning coincides with classical reasoning, i.e., reasoning over arbitrary, unrestricted models. Then, we analyze the description logics DL-Lite_F and DL_Lite_A. Differently from DL-Lite_R, in such logics finite model reasoning does not coincide with classical reasoning. To solve satisfiability and query answering over finite models in these logics, we define techniques which reduce polynomially both the above reasoning problems over finite models to the corresponding problem over arbitrary models. Thus, for all the DL-Lite languages considered, the good computational properties of satisfiability and query answering under the classical semantics also hold under the finite model semantics. Moreover, we have effectively and easily implemented the above techniques, extending the DL-Lite reasoner QuOnto with support for finite model reasoning.

@inproceedings{rosati2008finite, author = {Rosati, Riccardo}, title = {Finite model reasoning in DL-Lite}, editor = {Hauswirth, Manfred and Koubarakis, Manolis and Bechhofer, Sean}, booktitle = {Proceedings of the 5th European Semantic Web Conference}, series = {LNCS}, publisher = {Springer Verlag}, address = {Berlin, Heidelberg}, year = {2008}, url = {http://data.semanticweb.org/conference/eswc/2008/papers/291}, keywords = {ontologies reasoning description computational logics complexity formal-languages-2}, abstract = {The semantics of OWL-DL and its subclasses are based on the classical semantics of first-order logic, in which the interpretation domain may be an infinite set. This constitutes a serious expressive limitation for such ontology languages, since, in many real application scenarios for the Semantic Web, the domain of interest is actually finite, although the exact cardinality of the domain is unknown. Hence, in these cases the formal semantics of the OWL-DL ontology does not coincide with its intended semantics. In this paper we start filling this gap, by considering the subclasses of OWL-DL which correspond to the logics of the DL-Lite family, and studying reasoning over finite models in such logics. In particular, we mainly consider two reasoning problems: deciding satisfiability of an ontology, and answering unions of conjunctive queries (UCQs) over an ontology. We first consider the description logic DL-Lite_R and show that, for the two above mentioned problems, finite model reasoning coincides with classical reasoning, i.e., reasoning over arbitrary, unrestricted models. Then, we analyze the description logics DL-Lite_F and DL_Lite_A. Differently from DL-Lite_R, in such logics finite model reasoning does not coincide with classical reasoning. To solve satisfiability and query answering over finite models in these logics, we define techniques which reduce polynomially both the above reasoning problems over finite models to the corresponding problem over arbitrary models. Thus, for all the DL-Lite languages considered, the good computational properties of satisfiability and query answering under the classical semantics also hold under the finite model semantics. Moreover, we have effectively and easily implemented the above techniques, extending the DL-Lite reasoner QuOnto with support for finite model reasoning.} }

%0 = inproceedings %A = Rosati, Riccardo %B = Proceedings of the 5th European Semantic Web Conference %C = Berlin, Heidelberg %D = 2008 %I = Springer Verlag %T = Finite model reasoning in DL-Lite %U = http://data.semanticweb.org/conference/eswc/2008/papers/291