Search

This is the last of the three deliverables resulting from Task 2.3, Task 2.4 and Task 2.5. This set of deliverables presents the definition and implementation of the Cloud Certification Language which encompasses three major phases: 1) the encoding of requirements of cloud certification schemas – written in natural language -- in a Controlled Natural Language (CNL), so called MEDINA CNL; 2) the editing of the requirements in MEDINA CNL through an editor tool; and 3) the mapping of the CNL requirements to a domain specific language (DSL).

This is the second of the three deliverables resulting from Task 2.3, Task 2.4 and Task 2.5. This set of deliverables will present the definition and implementation of the Cloud Certification Language which encompasses three major phases: 1) the encoding of requirements of cloud certification schemas – written in natural language -- in a Controlled Natural Language (CNL), so called MEDINA CNL; 2) the editing of the requirements in MEDINA CNL through an editor tool; 3) the mapping of the CNL requirements to a domain specific language (DSL).

Pre-Print of the paper that will be published in the CEUR proceedings

This whitepaper focuses on the description of the software and hardware architecture of the MEDINA framework, which has been designed and implemented during the first 18 months of the EU MEDINA project. First, the report briefly introduces the problem of trustworthiness in cloud services in the EU, the continuous monitoring of cloud services and the certification issues. The MEDINA project aims to provide CSPs with a tool that allows them to audit and certify of Cloud Services in an automated and near real-time manner. The continuous certification proposed in MEDINA is aligned to the EUCS framework. The paper describes the principal concepts and ideas leading to the MEDINA framework, namely: the definition of a metrics catalogue based in EUCS; the implementation of a machine-readable language to allow the interpretation of natural-language specifications; the automated and continuous evidence collection and assessment; and the managing of digital evidences for accountability. Then, an overview of MEDINA is presented, which includes the roles, workflows and tasks related to the framework, as they have been defined during the first half of the project. Three are the main roles taking part in the basic workflow: the Compliance Manager, the Control Owner and the External Auditor. Seven workflows are defined, that compose the whole tasks that the users of MEDINA have to undertake to achieve their goals. The workflows comprise the preparation of the components, the definition of the target of certification (ToC), the deployment of the framework on the ToC, the self-assessment, the compliance assessment, and the maintenance and report of the certificate. This part is completed with a diagram of the system architecture, that decomposes the framework in components that are dedicated to specific tasks, and that collaborate to provide the explained functionality. Finally, the components of the MEDINA framework are described. The components in which we have divided MEDINA are the Catalogue of Controls & Security Schemes; the components that deal with the NLP (Natural Language Processing) techniques; the Risk Assessment and Optimisation Framework; the Continuous Evaluation; The Orchestrator and the Trustworthiness System; the Evidenced Collection and Security Assessment; the Certificate management System; and the Integrated User Interface that works as a wrapper of the rest of components for the final user. The functionality of each component is briefly presented, showing the interactions with the rest of the framework. It is not the objective of this white paper to present a deep view into the technical issues tackled by each of the components, but a birds-eye view that allows the reader to grasp a general vision of what MEDINA tries to offer. In this sense, some screenshots of the prototypes developed in the project are also included throughout the text, which can help the reader to get an idea of the current state of development of the tools.

The problem analyzed in this paper deals with the classification of Internet traffic. During the last years, this problem has experienced a new hype, as classification of Internet traffic has become essential to perform advanced network management. As a result, many different methods based on classical Machine Learning and Deep Learning have been proposed. Despite the success achieved by these techniques, existing methods are lacking because they provide a classification output that does not help practitioners with any information regarding the criteria that have been taken to the given classification or what information in the input data makes them arrive at their decisions. To overcome these limitations, in this paper we focus on an “explainable” method for traffic classification able to provide the practitioners with information about the classification output. More specifically, our proposed solution is based on a multi-objective evolutionary fuzzy classifier (MOEFC), which offers a good trade-off between accuracy and explainability of the generated classification models. The experimental results, obtained over two well-known publicly available data sets, namely, UniBS and UPC, demonstrate the effectiveness of our method.

In this paper, we introduce the rationale, goals, and structure of Presente Digitale, an ambitious Italian project for the realization of an online education system on Digital Culture. Presente Digitale is dedicated to teachers and their students. Its aim is to offer paths of reflection closely linked to the digital world, which can respond to the current needs of schools and society about Digital Culture and helps in effectively convey this knowledge. This will give young people the opportunity to integrate their prior knowledge with a ‘teacher’s points of view’, and, mostly important, to be able to take a safe and solid path when approaching the labour market. The courses are free of charge and available online, making the participation really flexible.

Bot accounts are computer algorithms populating our online ecosystem. They often feature goals that are far from being benign: distribution of spam, inflated sponsorship of public characters, biased influence on the audience's judgements. The endless game of cat and mouse between Academia and bot developers proves that bots evolve over time, imitating human online behaviour and making themselves undetectable even by state-of-art detection tools. This peculiarity inspires the current work: the adoption of text mining tools to evaluate if it is possible to infer typology and purpose of bots from the feelings and interests expressed in their posts. The experiments, carried out on a large corpus of Twitter data, show a connection between the two considered dimensions (feeling and interest) and the characteristics of the evolutionary waves of accounts.

Nowadays, a huge amount of data are generated, often in very short time intervals and in various formats, by a number of different heterogeneous sources such as social networks and media, mobile devices, internet transactions, networked devices and sensors. These data, identified as Big Data in the literature, are characterized by the popular Vs features, such as Value, Veracity, Variety, Velocity and Volume. In particular, Value focuses on the useful knowledge that may be mined from data. Thus, in the last years, a number of data mining and machine learning algorithms have been proposed to extract knowledge from Big Data. These algorithms have been generally implemented by using ad-hoc programming paradigms, such as MapReduce, on specific distributed computing frameworks, such as Apache Hadoop and Apache Spark. In the context of Big Data, fuzzy models are currently playing a significant role, thanks to their capability of handling vague and imprecise data and their innate characteristic to be interpretable. In this work, we give an overview of the most recent distributed learning algorithms for generating fuzzy classification models for Big Data. In particular, we first show some design and implementation details of these learning algorithms. Thereafter, we compare them in terms of accuracy and interpretability. Finally, we argue about their scalability.

Nowadays, companies and enterprises are more and more incline to exploit the pervasive action of on-line social media, such as Facebook, Twitter and Instagram. Indeed, several promotional and marketing campaigns are carried out by concurrently adopting several social medial channels. These campaigns reach very quickly a wide range of different categories of users, since many people spend most of their time on on-line social media during the day. In this work, a Decision Support System (DSS) is presented, which is able to efficiently support companies and enterprises in managing promotional and marketing campaigns on multiple social media channels. The proposed DSS continuously monitors multiple social channels, by collecting social media users’ comments on promotions, products, and services. Then, through the analysis of these data, the DSS estimates the reputation of brands related to specific companies and provides feedbacks about a digital marketing campaign. The core of the proposed DSS is a Sentiment Analysis Engine (SAE), which is able to estimate the users’ sentiment in terms of positive, negative or neutral polarity, expressed in a comment. The SAE is based on a machine learning text classification model, which is initially trained by using real data streams coming from different social media platforms specialized in user reviews (e.g., TripAdvisor). Then, the monitoring and the sentiment classification are carried out on the comments continuously extracted from a set of public pages and channels of publicly available social networking platforms (e.g., Facebook, Twitter, and Instagram). This approach is labeled as cross-source sentiment analysis. After some discussions on the design and the implementation of the proposed DSS, some results are shown about the experimentation of the proposed DSS on two scenarios, namely restaurants and consumer electronics online shops. Specifically, first the application of effective sentiment analysis models, created relying on user reviews is discussed: the models achieve accuracies higher than 90%. Then, such models are embedded into the proposed DSS. Finally, the results of a social listening campaign are presented. The campaign was carried out by fusing data streams coming from real social channels of popular companies belonging to the selected scenarios.

In this paper, we propose and test an approach based on regression models, to predict the review score of an item, across different reviewer categories. The analysis is based on a public dataset with more than 2.5 million hotel reviews, belonging to five specific reviewers’ categories. We first compute the relation between the average scores associated with the different categories and generate the corresponding regression model. Then, the extracted model is used for prediction: given the average score of a hotel according to a reviewer category, it predicts the average score associated with another category.

Social network sites, also called micro-blogging services (e.g., Twitter, Facebook, Google+), have spread in recent years, becoming a new kind of real-time information channel. Their popularity stems from the characteristics of portability thanks to several social networks applications for smartphones and tablets, easiness of use, and real-time nature. People intensely use social networks to report (personal or public) real-life events happening around them or simply to express their opinion on a given topic, through a public message. Social networks allow people to create an identity and let them share it in order to build a community. In this talk, we first discuss how social networks and media platforms can be used as sources of information for the detection of events, such as traffic congestion, incidents, natural disasters (earthquakes, storms, fires, etc.), or other events. Then, we show how to extract the “sentiment” hidden behind messages and comments that users post and share on social networks. The users' sentiment may be exploited for profiling the reputation of commercial activities and brands, for estimating the results of elections or of marketing campaigns, and in general for evaluating the general humor of a community with respect a specific topic.

How to deploy and manage, in an efficient and adaptive way, complex applications across multiple heterogeneous cloud platforms is one of the problems that have emerged with the cloud revolution. In this paper we present context, motivations and objectives of the EU research project SeaClouds, which aims at enabling a seamless adaptive multi-cloud management of complex applications by supporting the distribution, monitoring and migration of application modules over multiple heterogeneous cloud platforms. After positioning SeaClouds with respect to related cloud initiatives, we present the SeaClouds architecture and discuss some of its aspect, such as the use of the OASIS standard TOSCA and the compatibility with the OASIS CAMP initiative.

A multi-objective evolutionary fuzzy rule selection process extracts a subset of fuzzy rules from an initial set, by applying a multi-objective evolutionary algorithm. Two approaches can be used to determine the number of terms (i.e. the granularity) associated with the linguistic variables that appear in the rules: a pre-established single granularity can be chosen, or a multiple granularities approach can be preferred. The latter favors a reduction in the number of extracted rules, but it also brings to a possible loss of interpretability. To prevent this problem, suitable granularities can be determined by applying automatic techniques before the initial rule generation process. In this contribution, we investigate how the application of a single granularity learning approach influences the performance of fuzzy associative rule-based classifiers. The aim is to reduce the complexity of the obtained models, trying to maintain a good classification ability.

When considering data sets characterized by a large number of instances, the computational time required to apply Genetic Algorithms for generating Fuzzy Rule-Based Classifiers increases considerably, mainly due to the fitness evaluation. Another important problem associated to these kinds of data sets is an undesired increase of the obtained model complexity.