Search

Part 2: Full Papers; International audience; Annotations obtained by Cultural Heritage institutions from the crowd need to be automatically assessed for their quality. Machine learning using graph kernels is an effective technique to use structural information in datasets to make predictions. We employ the Weisfeiler-Lehman graph kernel for RDF to make predictions about the quality of crowdsourced annotations in Steve.museum dataset, which is modelled and enriched as RDF. Our results indicate that we could predict quality of crowdsourced annotations with an accuracy of 75 %. We also employ the kernel to understand which features from the RDF graph are relevant to make predictions about different categories of quality.

International audience; The analysis of security incidents and frauds has shown that several vulnerabilities of IT systems are due to loopholes in the policies and procedures adopted by organizations as well as in their structure. Organizations have thus to address security and dependability issues by analyzing their organizational setting. In this paper, we present a methodology to support the deployment of Security & Dependability patterns according to their position in the Enterprise Architecture and the underlying system infrastructures. The methodology discriminates the pattern deployment process between recommendations and guidelines. Recommendations concretize the deployment with refined software and/or hardware related patterns, whereas guidelines specify the organizational patterns in terms of the system-to-be, proposing human-resource and/or policy solutions. To make the discussion more concrete, we illustrate the framework with a case study on an emergency scenario within a remote healthcare system.

In this paper we study probable innocence, a notion of probabilistic anonymity provided by protocols such as Crowds. The authors of Crowds, Reiter and Rubin, gave a definition of probable innocence which later has been interpreted by other authors in terms of the probability of the users from the point of view of the observer. This formalization however does not seem to correspond exactly to the property that Reiter and Rubin have shown for Crowds, the latter, in fact, is independent from the probability of the users.We take the point of view that anonymity should be a concept depending only on the protocol, and should abstract from the probabilities of the users. For strong anonymity, this abstraction leads to a concept known as conditional anonymity. The main goal of this paper is to establish a notion which is to probable innocence as conditional anonymity is to strong anonymity. We show that our definition, while being more general, corresponds exactly to the property that Reiter and Rubin have shown for Crowds, under specific conditions. We also show that in the particular case that the users have uniform probabilities we obtain a property similar to the definition of probable innocence given by Halpern and O’Neill.