Istituto di Scienza e Tecnologie
dell’Informazione “A. Faedo”

Speakers:

Michele Coscia

Referent:

Andrea Esuli

In the last few years many real-world networks have been found to show a so-called community structure organization. Much effort has been devoted in the literature to develop methods and algorithms that can efficiently highlight this hidden structure of the network, traditionally by partitioning the graph. Since network representation can be very complex and can contain different variants in the traditional graph model, each algorithm in the literature focuses on some of these properties and establishes, explicitly or implicitly, its own definition of community. According to this definition it then extracts the communities that are able to reflect only some of the features of real communities. The aim of this survey is to provide a manual for the community discovery problem. Given a meta definition of what a community in a social network is, our aim is to organize the main categories of community discovery based on their own definition of community. Given a desired definition of community and the features of a problem (size of network, direction of edges, multidimensionality, and so on) this review paper is designed to provide a set of approaches that researchers could focus on.

NOTE: This seminar is the sixth one of the series of six seminars presented by the winners of the prize "Young researchers ISTI 2013". Michele Coscia placed second in the Young Researcher category.

Speakers:

Anastasios Noulas (Cambridge University)

Referent:

Fabrizio Silvestri

In this talk I will be sharing my experience of mining, analysing and modeling user mobility in location-based social networks in the past couple of years. We will initially present an agent based modeling approach that captures fundamental properties of human urban movement in 34 metropolitan cities. The work that is based on a dataset comprised of millions of Foursquare user check-ins will shed light on the role of geography and the spatial density of settlements in human mobility. Subsequently, we will consider two different mobile application scenarios. First we will formulate a mobile recommendation task where the goal is to accurately new places a user will visit in future time periods. We shall see how a random walk on a graph that connects users with places can offer good recommendations in an extremely sparse data context. Next, we shall formulate an even more challenging problem where we try to predict the next check-in of a user in real time. Exploiting a supervised learning approach that employs synchronously multiple layers of mobility data, we will show that accurate predictions of user whereabouts can be attained if models are trained appropriately. Finally, we will discuss on the feedback loop between abstract models and applications and how mutual benefits emerge in two seemingly distant areas of research in human mobility.

Biografia: Anastasios Noulas is a PhD candidate in the Computer Laboratory at Cambridge University. He holds an MEng in Computer Science (2009) from University College London. His research interests are focused on the analysis and modelling of human movement and geographic social networks, with techniques that span the areas of Complex Systems, Data Mining and applied Machine Learning. In 2012 he joined Telefonica Research in Spain for a research fellowship participating in a project on Smart Cities that exploited Call Detail Records and data sourced from location-based services to infer activities in urban environments. As of 2013 he has joined the EPSRC project GALE that aims to the development of the next generation of mobile recommender systems and urban neighbourhood modelling for the provision of local knowledge to remote visitors of an area. More details are available http://www.cl.cam.ac.uk/~an346/

Speakers:

Cristina Muntean

Referent:

Andrea Esuli

Touristic applications stirred an increased interest in the last years due to the intense use of mobile devices and location based applications. Our outlook on this matter is directed towards the next point of interest (PoI) prediction task. We tackle the problem of predicting the “next” geographical position of a tourist, given her history (i.e., the prediction is done accordingly to the tourist’s current trail) by means of supervised learning techniques.
We test our methods on three datasets built using geo-tagged pictures downloaded from Flickr, each collection corresponding to a popular touristic area. We adopt two popular Machine Learning methods, namely Gradient Boosted Regression Trees and Ranking SVM for learning to rank the next PoI, on the basis of an object space represented by a multi-dimensional feature vector, specifically designed for tourism related data. We define a set of 68 different features, broadly classified into two main categories, namely “Session” and “PoI”. Session features are meant to model the tourist behavior and capture concepts like groups of PoIs visited, distances among PoIs and other characteristics of a user session (trail). On the other hand, PoI features model the characteristics of a candidate PoI, also taking into account the past activities of the tourist.
We propose a thorough comparison of several methods that are considered state-of-the-art in touristic recommender and trail prediction systems (WhereNext, Random Walk with Restart), as well as a strong popularity baseline. As experiments show, the methods we propose constantly outperform, with up to 300% in terms of prediction accuracy, our baselines and provide strong evidence of the performance and robustness of our solutions.

NOTE: This seminar is the fifth one of the series of six seminars presented by the winners of the prize "Young researchers ISTI 2013". Cristina Muntean placed third in the PhD student category.

Speakers:

Roberta Lanziani (IMT Lucca)

Referent:

Mieke Massink

In this talk we investigate the use of Central Limit Approximation of Continuous Time Markov Chains to verify collective properties of large population models, describing the interaction of many similar individual agents. More precisely, we specify properties in terms of individual agents by means of deterministic timed automata with a single global clock (which cannot be reset), and then use the Central Limit Approximation to estimate the probability that a given fraction of agents satisfies the local specification.

Speakers:

Laura Nenzi (IMT Lucca)

Referent:

Mieke Massink

We present a new approach for the design of a synthetic biological circuit whose behavior is specified in terms of signal temporal logic (STL) formulae. We first show how to characterize with STL formulae the input/output behavior of biological modules miming the classical logical gates (AND, NOT, OR). Hence, we provide the regions of the parameter space for which these specifications are satisfied. Given a STL specification of the target circuit to be designed and the networks of its constituent components, we propose a methodology to constrain the behavior of each module, then identifying the subset of the parameter space in which those constraints are satisfied, providing also a measure of the robustness for the target circuit design. This approach, which leverages recent results on the quantitative semantics of Signal Temporal Logic, is illustrated by synthesizing a biological implementation of an half-adder.

Speakers:

Alessio Ferrari

Referent:

Andrea Esuli

We present a novel approach for pragmatic ambiguity detection in natural language (NL) requirements specifications defined for a specific application domain. Starting from a requirements specification, we use a Web-search engine to retrieve a set of documents focused on the same domain of the specification. From these domain-related documents, we extract different knowledge graphs, which are employed to analyse each requirement sentence looking for potential ambiguities. To this end, an algorithm has been developed that takes the concepts expressed in the sentence and searches for corresponding "concept paths" within each graph. The paths resulting from the traversal of each graph are compared and, if their overall similarity score is lower than a given threshold, the requirements specification sentence is considered ambiguous from the pragmatic point of view. A proof of concept is given throughout the presentation to illustrate the soundness of the proposed strategy.

NOTE: This seminar is the fourth one of the series of six seminars presented by the winners of the prize "Young researchers ISTI 2013". Alessio Ferrari placed first in the Young researcher category.

Speakers:

Simon Davies ( LSE Enterprise, a wholly owned subsidiary of The London School of Economics and Political Science, London, UK)

Referent:

Diego Latella

Europe - along with much of the developing world - is ready to implement a new and robust regime of data protection. The IT industry is carefully considering the challenges and the opportunities that these new legal regulations may create. Data protection provides a pillar of trust necessary to nurture emerging services and products. Even so, concern has been expressed that the new rules could hinder innovation and create barriers to design and engineering. Some companies believe the regulatory bar has now been set too high and that data protection will create substantial problems. In this talk veteran privacy expert Simon Davies discusses whether the proposed rules have struck the right formula.

Simon Davies is the Founder of Privacy International and Associate Director of LSE Enterprise. He has been a Visiting Fellow in Law at both the University of Greenwich and the University of Essex, and spent 13 years at LSE, where he taught the groundbreaking MSc Masters course in "Privacy & Data Protection". He is also co-director of LSE’s Policy Engagement Network. Simon Davies is widely acknowledged as one of the most influential data protection and internet rights experts in the world and is a pioneer of the international privacy arena. His work in consumer rights and technology policy has spanned over 25 years and has directly influenced the development of law and public policy in more than 40 countries. He has advised a wide range of corporate, government and professional bodies, including the United Nations High Commissioner for Refugees. Recently, Simon Davies has been tasked by cross-party rapporteurs of the European Parliament to conduct a wide-ranging external assessment of the European Commission's proposed reforms to the EU data protection framework. He brings a unique interface with global stakeholders, from major international corporations to government and civil society.

Nota: Martedi 21 maggio, cioè il giorno prima del seminario, alle ore 15, in Aula Faedo, verrà proiettata la registrazione di una puntata del programma Cyborg City della CNN che comprende, tra l'altro, un'intervista a Simon Davies.

Speakers:

Giacomo Berardi

Referent:

Andrea Esuli

Suppose an organization needs to classify a set D of textual documents, and suppose that D is too large to be classified manually, so that resorting to some form of automated text classification (TC) is the only viable option. Suppose also that the organization has strict accuracy standards, so that the level of effectiveness obtainable via state-of-the-art TC technology is not sufficient. In this case, the most plausible strategy to follow is to classify D by means of an automatic classifier F, and then to have a human editor inspect the results of the automatic classification, correcting misclassifications where appropriate. The human annotator will obviously inspect only a subset D' of D (since it would not otherwise make sense to have an initial automated classification phase). We call this scenario Semi-Automated Text Classification (SATC). An automated system can support this process by ranking the automatically labelled documents in a way that maximizes the expected increase in effectiveness that derives from inspecting D'. An obvious strategy is to rank D so that the documents that F has classified with the lowest confidence are top-ranked. In this work we show that this strategy is suboptimal. We develop a new utility-theoretic ranking method based on the notion of inspection gain, defined as the improvement in classification effectiveness that would derive by inspecting and correcting a given automatically labelled document. We also propose a new effectiveness measure for SATC-oriented ranking methods, based on the expected reduction in classification error brought about by partially inspecting a list generated by a given ranking method.
We report the results of experiments showing that, with respect to the baseline method above, and according to the proposed measure, our ranking method can achieve substantially higher expected reductions in classification error.

NOTE: This seminar is the third one of the series of six seminars presented by the winners of the prize "Young researchers ISTI 2013". Giacomo Berardi placed first in the PhD student category.

Speakers:

Midhat Ali (University of Camerino)

Referent:

Andrea Polini

Platform providers establish marketplace ecosystems to sell to end users services and applications running on the platform. Very prominent examples of these marketplaces are the Google Play Store and Apple AppStore, among others. Application Developers can use these marketplaces to offer services and applications to end users.
Advanced Security Service cERTificate (ASSERT) for SOA (ASSERT4SOA) is a Framework designed to associate certificates of Security properties with the applications and services. This framework offers services like verification of ASSERTs and matching those ASSERTs to Security Properties specified in an appropriate query language. ASSERT Enabled Marketplace is a prototype to qualitatively evaluate and illustrate the use of ASSERT4SOA Framework in a marketplace which sells business applications, which have certificates issued for certain security properties. This marketplace is designed based on a typical user story of buying business applications based on the organization’s security requirements, comparing and choosing the best option based on the requirements.

Speakers:

Allyson Dale (PhD student, Department of Psychology, Trent University, Ontario, Canada)

Referent:

Umberto Barcaro

Dream content and discovery have been studied comparing soldiers with operational experience in Afghanistan with an age matched sample of male civilians. Soldiers' dreams contained a higher frequency of aggression, threat, and military imagery, consistent with the continuity hypothesis.Dream interpretation for soldiers led to discovery about specific events from tours overseas and about aggressive behaviours.Further research compared the dream content of students, soldiers, and heavy gamers and results revealed significant differences in dream imagery for soldiers when compared to both students and gamers.

Speakers:

Anthony Murkar (Masters student, Department of Psychology, Trent University, Ontario, Canada)

Referent:

Umberto Barcaro

The seminar presentation will discuss how dreams relate to memory. It is not fully understood how dreams are generated by the brain, but recent research has uncovered relationships between sleep, dreams, and memory function (Peigneux et al., 2004; Wamsley et al., 2010). Implications for understanding how brain activity relates to dreaming are discussed.

Speakers:

Nicolle Miller (Masters level student, Department of Psychology, Trent University, Ontario, Canada)

Referent:

Umberto Barcaro

The current study examined the effects of meditation on waking day depression levels (BDI), waking day trait anxiety levels (BAI-T) and dream imagery in University students. The Storytelling Method was used to conduct dream interpretations, using word associations and story narrative. This seminar will address the results, which are consistent with past research, as well as the implications and applications of dream work and meditation in clinical and applied practice.

Speakers:

Rossano Venturini

Referent:

Andrea Esuli

Given a set of integer keys from a bounded universe along with associated data, the dictionary problem asks to build a data structure able to answer efficiently two queries: membership and retrieval.
Membership has to tell whether a given element is in the dictionary or not; Retrieval has to return the data associated with the searched key.
This seminar will describe a recent result presented at ACM-SIAM SODA 2013 which provides time and space optimal solutions for three well-established relaxations of this basic problem: (Compressed) Static functions, Approximate membership and Relative membership.

NOTE: This seminar is the second one of the series of six seminars presented by the winners of the prize "Young researchers ISTI 2013". Rossano Venturini placed third in the Young researcher category.

Speakers:

Richard Connor (University Of Strathclyde, Glasgow, Uk)

Referent:

Fausto Rabitti

Over the years a number of competing models have been introduced attempting to solve the central IR problem of ranking documents given textual queries. These models, however, tend to require the inclusion of heuristics and the estimation of collection-specific parameter values in order to be effective. We define a new model that we do not believe has yet been explored. In terms of the categorisation of IR models, it is a probabilistic model and has no term inter-dependencies, thus allowing calculation from inverted indices. It is based upon a simple core hypothesis, directly calculating a ranking score in terms of probability theory and does not require the estimation of any parameters. We show initial tests in comparison with a number of standard baseline IR models, and show that the new model is at least credible, often outperforming the Language Model with Dirichlet smoothing.

Our contributions are twofold: first, we believe the new model is worthy of further investigation and that its performance could be improved significantly; and secondly, we believe the observation that the Jensen-Shannon metric can be evaluated over inverted indices in a sparse space is also more generally applicable.

Speakers:

Giulio Rossetti

Referent:

Giulio Rossetti

Community discovery in complex networks is an interesting problem with a number of applications, especially in the knowledge extraction task in social and information networks. However, many large networks often lack a particular community organization at a global level. In these cases, traditional graph partitioning algorithms fail to let the latent knowledge embedded in modular structure emerge, because they impose a top-down global view of a network. DEMON is a simple local-first approach to community discovery, able to unveil the modular organization of real complex networks. This is achieved by democratically letting each node vote for the communities it sees surrounding it in its limited view of the global system, i.e. its ego neighborhood, using a label propagation algorithm; finally, the local communities are merged into a global collection.

NOTE: This seminar is the first one of the series of six seminars presented by the winners of the prize "Young researchers ISTI 2013". Giulio Rossetti placed second in the PhD student category.