Intelligenza Artificiale

Statistical Relational Learning and Probabilistic Inductive Logic Programming are two emerging fields that use representation languages able to combine logic and probability. In the field of Logic Programming, the distribution semantics is one of the prominent approaches for representing uncertainty and underlies many languages such as ICL, PRISM, ProbLog and LPADs. Learning the parameters for such languages requires an Expectation Maximization algorithm since their equivalent Bayesian networks contain hidden variables. EMBLEM (EM over BDDs for probabilistic Logic programs Efficient Mining) is an EM algorithm for languages following the distribution semantics that computes expectations directly on the Binary Decision Diagrams that are built for inference. In this paper we present experiments comparing EMBLEM with LeProbLog, Alchemy, CEM, RIB and LFI-ProbLog on six real world datasets. The results show that EMBLEM is able to solve problems on which the other systems fail and it often achieves significantly higher areas under the Precision Recall and the ROC curves in a similar time.

This work focuses on the topic of cultural stories, about which we propose a methodology for the annotation of media that presents historical and cultural contents, the composition of the stories, and the production of multimedia presentations of aggregated media. Specifically, the research consists in the definition of an annotation language for events in which time is represented in explicit form (qualitative and quantitative). The logical representation of the stories has allowed for building specialized modules to facilitate different types of events analysis associated with the media regarding the checking of connectivity and consistency, causality, and the discovery of temporal relations. In this work, concepts belonging to Film Theory, like the notions of plot, fabula, flashback and flashforward, were adopted for the construction of a story. Specific diagrams for the comparison of different ways (plots) in which a story can be told were proposed. Among the tools developed in this research, we present the algorithms implemented for the production of the code for visualizing multimedia stories, starting from the logical representation of annotated events.

Affective computing is receiving increasing attention in many sectors, ranging from advertisement to politics. This work, set in a Social Semantic Web framework, presents ArsEmotica, an application software for associating the predominant emotions to artistic resources of a social tagging platform. Our aim is to extract a rich emotional semantics (i.e. not limited to a positive or a negative reception) of tagged resources through an ontology driven approach. This is done by exploiting and combining available computational and sentiment lexicons with an ontology of emotional categories. The information sources we rely upon are the tags by which users annotated resources, that are available through the ArsMeteo platform, and the ontology OntoEmotion, that was enriched by means of our tool with over four hundred Italian emotional words referring to the about eighty-five emotional concepts of the ontology. Tags directly referring to ontological concepts are identified, while potentially affective tags can be annotated by using the ontology, thanks to the spontaneous intervention of the users, in a pure Web 2.0 approach. Finally, the tagged artworks are related with the emerging predominant emotions. A user study involving the ArsMeteo community was conducted in order to evaluate the ArsEmotica outcome, for what concerns the emotions automatically associated by the system to the artworks.

Pricing Web services with different Quality of Service (QoS) classes over the Internet is a challenging scientific problem. Finding the optimal price per QoS class allows each provider to maximize its revenue and to solve the capacity planning problem. The most works presented in the literature focus on a monopolistic setting where one service provider sells one or more services of different classes to a number of customers. The Service Level Agreement (SLA) is the most commonly used contract scheme, whereby providers announce their QoS profiles specifying a number of QoS classes and the related prices. The most results available in the literature to find the optimal QoS profile of a provider do not take into account the QoS profiles announced by other providers. However, in real–world settings, services sale competition among service providers is common and cannot be neglected. In this paper, we design a game–theoretical model for Web service markets with a single service characterized by response time and price by extending the Bertrand–Edgeworth's oligopoly, and we provide a polynomial–time algorithm to find a Nash equilibrium with two providers and an arbitrary number of customers.

Probabilistic Inductive Logic Programming and Statistical Relational Learning are families of techniques that are exploited in Machine Learning applications to perform advanced tasks in several domains. Every day the size and complexity of such problems increases and advanced, expressive and efficient tools are needed to successfully solve them. The literature proposes several algorithms to cope with these problems, each of them with its own quirks and perks. Among various solutions, Logic Programming with Annotated Disjunctions (LPAD) is one of the more attractive formalisms, thanks to the expressiveness and readability of its language. Unfortunately, its most advanced implementations are lacking efficient features and techniques that have been introduced for other formalisms, such as ProbLog. In this work, after introducing LPADs and an inference algorithm for computing the probability of a query, we investigate four different approximated algorithms, inspired by similar work done in ProbLog. In particular, we present each algorithm and we evaluate its performances on real and artificial datasets. The results show that our approaches have performances that are usually in line with ProbLog. The Monte Carlo algorithm, however, has performances that are better than the exact approach in terms of both the maximum size of the problems and the execution time, with a neglectable loss in the accuracy of the result.

The greater and greater availability of services over the web motivates the growing interest in techniques that facilitate their re-use. A web service specification can be quite complex, including various operations and message exchange patterns. In this work, we propose a rule-based declarative representation of services, and in particular of WSDL operations, that enables the application of techniques for reasoning about actions and change, that are typical of agent systems. The representation allows reasoning on choreography roles and on possible role players, to the aim of selecting services which match in a flexible way with the specification. Flexible matches are an important tool that allows web service re-use but the proposals in the literature do not guarantee the preservation of those goals, that can be proved over the role specification. We show how to enrich various well-known matches so as to produce substitutions that preserve goals and that do not require service rollback. We also discuss the problem of the joint achievement of the individual goals of a group of choreography role players.