GDRI "Web Science" Meeting, Toulouse 10-12 September 2014

AmphithŽ‰tre Jacques Herbrand, IRIT, UniversitŽ Paul Sabatier

Workshop Agenda

Wednesday September 10




An Algebra of Lightweight Ontologies
Marco Antonio Casanova, PUC Rio




Inferring same-as facts from Linked Data: an iterative import-by-query approach
Marie-Christine Rousset, LIG Grenoble


Data Interlinking based on Semantic Graphs
Bernardo Pereira Nunes, PUC Rio


Complex Ontology Correspondences for Query Patterns Rewriting in the context of a Question-Answering System over Linked Data, Cassia Trojahn, IRIT Toulouse




Analysis and verification of data-centeric web services
Farouk Toumani, LIMOS Clermont-Ferrand


Exploring Semi Structured Data
Daniel Schwabe, PUC Rio


Web archiving [slides]
StŽphane Ganarski, LIP6 Paris


Linked data, Complex Networks and a Web Infrastructure for eBiology
AndrŽ Santanch, UNICAMP S‹o Paulo




Gime a lever and a fulcrum for ATA (Automatic Text Analysis) [slides]
Marc El-Bze, Agorantic Avignon


Sentiment Recognition in Text Through Affective Computing
Geraldo Xexeo, UFRJ Rio


Workflow Provenance and Data Quality Assessment in WebLab
ClŽment Caron, LIP6 Paris


Approaches to the dataset interlinking recommendation problem
Giseli Rabello Lopes, PUC Rio


Dinner, "La table de William", Toulouse,

Thursday September 11


Social media search: leveraging user generated content
Mohand Boughanem, IRIT Toulouse




a) Representation and processing of some dynamic aspects of trajectoriesand (b) Content diffusion in social networks.
Eric Gaussier et Herve Martin, LIG Grenoble


Big Social Data: Analyzing and extracting knowledge from social data in web [slides]
Jonice Oliveira, UFRJ Rio


Brazilian research in Scientometrics: history, trends and challenges
Jacqueline Leta, UFRJ Rio


Scientometric Studies of Academic Writing [slides]
Guillaume Cabanac, IRIT Toulouse




Trajectory Enrichment using Linked Data
JosŽ Antonio de Macedo, UFC Rio


Spatial Big Data Handling - Examples in Astronomy and Transport
Karine Zeitouni, PRISM Versailles


Aggregated Search in Graph Database: A Relational-Based Approach
Mohand-Said Hacid, LIRIS Lyon


Representation Learning for image understanding
Nicolas Thome, Lip6




Collaborative and Open Government Systems
Renata Araœjo, UNIRIO


Tr@nspolo : Focus on the analysis of the staging of everyday political work between political and computer sciences [slides]
Guillaume Marrel, Agorantic Avignon


Developing expertise in the design of digital democracy : a comparative and transdisciplinary approach (France-QuŽbec-Brazil) [slides]
Magali Nonjon, Agorantic Avignon


Citizen-Centric Business Process [slides]
Yahia Taher, PRISM Versailles Saint Quentin


Dinner, "Le Moai", Toulouse,

Friday September 12


Profile Diversity for Query Processing using Users Recommendations
Esther Pacitti, INRIA & LIRMM Montpellier


Analyzing Social Presence and Willingness in Online Social Networks
Sean Siqueira, UNIRIO


Museum, web and equipped visitors. Interactions and appropriation of knowledge by participatory tools
Florence Andreacola, Agorantic Avignon


Towards more targeted recommendations in folksonomies
Engelbert Mephu Nguifo, LIMOS Clermont-Ferrand




Cooperation tools between France (Europe) and Brazil (south America)
Antonia Alcaraz, CNRS-ChargŽe de programme AmŽrique latine


Brainstorming (next steps, ...)







Wednesday September 10

Marco Antonio Casanova

Departamento de Inform‡tica     

Pontif’cia Universidade Cat—lica do Rio de Janeiro       

An Algebra of Lightweight Ontologies


Abstract: In this talk, we argue that certain ontology design problems are profitably addressed by treating ontologies as theories and by defining a set of operations that create new ontologies, including their constraints, out of other ontologies. We first show how to use the operations in the context of ontology reuse, how to take advantage of the operations to compare different ontologies, or different versions of an ontology, and how the operations may help design mediated schemas in a bottom up fashion. We then discuss how to compute the operations for lightweight ontologies, including the question of minimizing the set of constraints of a lightweight ontology.

Marie-Christine Rousset (joint work with Mustafa A Bakri and Manuel Atencia)

LIG, Grenoble


Inferring same-as facts from Linked Data: an iterative import-by-query approach


Abstract: In this work, we model the problem of data linkage in Linked Data as a reasoning problem on possibly decentralized data.

We describe a novel Import-by-Query algorithm that alternates steps of sub-query rewriting and of tailored querying the Linked Data cloud in order to import data as specific as possible for inferring or contradicting given target same-as facts.

Experiments conducted on real-world datasets have demonstrated the feasibility of this approach and its usefulness in practice for data linkage and disambiguation.

Cassia Trojahn

IRIT, Toulouse

Complex Ontology Correspondences for Query Patterns Rewriting in the context of a Question-Answering System over Linked Data


Abstract: We will present our work on complex ontology alignments in the task of query patterns rewriting. We apply this approach in SWIP, a system that allows for querying RDF data from natural language-based queries, hiding the complexity of SPARQL. SWIP is based on the use of query patterns that characterize families of queries and that are instantiated with respect to the initial user query expressed in natural language. However, these patterns are specific to the vocabulary used to describe the data source to be queried. For rewriting query patterns, we experiment ontology matching approaches in order to find (complex) correspondences between two ontologies describing data sources. We will as well give an overview on our current collaborations with Brazilian research labs on hierarchical relation extraction, ontology alignment visualisation, ontology debugging (using belief revision and formal ontology), and formal representation of roles and qualia in formal ontology.

Bernardo Pereira Nunes 

Departamento de Inform‡tica     

Pontif’cia Universidade Cat—lica do Rio de Janeiro


Data Interlinking based on Semantic Graphs


Abstract: One key feature of the Semantic Web lies in the ability to link related Web resources. However, while relations within particular datasets are often well-defined, links between disparate datasets and corpora of Web resources are rare. The increasingly widespread use of cross-domain reference datasets, such as Freebase and DBpedia for annotating and enriching datasets as well as documents, opens up opportunities to exploit their inherent semantic relationships to align disparate Web resources. In this talk, we will present a combined approach to uncover relationships between disparate entities and documents that exploits (a) graph analysis of reference datasets together with (b) entity co-occurrence on the Web with the help of search engines. The combined approach is adopted from social network theory to measure the connectivity between given entities in reference datasets as well as to identify connected Web resources. 



Farouk Toumani,

Limos, Clermont Ferrand




Analysis and verification of data-centric services


Abstract: Recently, the need of incorporating data as a first-class citizen in business protocols has been widely recognized and a number of research works has been carried out in this direction, laying the foundations of a data-centric approach to web services. Formal models used to describe data-centric services are essentially communicating guarded transitions systems in which transitions are used to model either messages exchanges between a service and its environment, or actions (i.e., read, write) over a global database shared among existing services.  The incorporation of data turns out to be very challenging since it makes service specifications infinite which leads, in most cases, to the undecidability of many verification problems. This talk addresses the problem of analysing specifications of data-centric Web service interaction protocols. It focuses more specifically on the problem of checking the simulation preorder in this context and discusses some related decidability and complexity issues.

Daniel Schwabe         

Departamento de Inform‡tica     

Pontif’cia Universidade Cat—lica do Rio de Janeiro


Exploring Semi Structured Data


Abstract: There has been a large growth of available semi-structured data on the Web, spurred both by governmental requirements for publishing public data, and by private sector, for various purposes. One such large initiative is the Linked Open Data Cloud. An increasingly important activity is to make sense of such published data, often exploring it as a prelude or as initial steps to perform some information processing task. Exploration is then a generalization of the traditional search task, as it involves other operations beyond finding specific information. The design and evaluation of exploratory frameworks is a complex, multi-disciplinary endeavor, with important challenges for both aspects. In this talk, we will argue the need to separate the conceptual exploratory operations users may carry out over semi-structured data from the particular interface designs used to give users access to such operations. We will illustrate the problems, and outline possible solutions for both conceptual operations and interface design aspects, and discuss how this separation of concerns allows more accurate evaluation of the relevant aspects of any proposed tool or framework that aims at supporting Explorations.

Stephane Gancarski

LIP6, Univ. Pierre Marie Curie

Web archiving


Abstract: The talk presents most of the research led at LIP6 on Web archiving, within the context of the CArtec french ARN project and the Scape European project. We first describe our approach for optimizing the temporal completeness the crawling of Web pages during the pages harvest. Then we present the issues related with accessing existing archives. Finally we focus on Web page segmentation, which is one of the original point of our approach.

AndrŽ Santanch       

Departamento de Sistemas de Informa‹o

Instituto de Computa‹o

Universidade Estadual de Campinas


Linked data, Complex Networks and a Web Infrastructure for eBiology


Abstract: Complex networks is a graph representation of discrete systems in terms of entities and relationships, with non-trivial topological features. It is a powerful approach to support scientific research, which can exploit the patrimony of CS studies in graphs as a scientific tool. The synergy among complex networks, Linked Data and social networks is still a wide field to explore, with several applications in eScience. The research topics presented in this research address this synergy in the eBiology context. It argues whether it is possible to conceive an infrastructure for eBiology, designed on top of a knowledge network, materialized in a graph database. Challenges in this infrastructure include how to design a topology-aware query language to support exploratory analysis, and how to exploit the latent semantics in socially produced content, combining the notion of dataspaces' "pay-as-you-go" integration with graphs, Linked Data and ontologies.




Marc El-Beze,

AGorantic, Univ. dÕAvignon


Gime a lever and a fulcrum for ATA (Automatic Text Analysis)


Abstract: When the extraction of information from a given set of texts is focused on opinion and/or thematic detection, it can be seen as a categorization problem, sometimes as a classification one. In the first case, a corpus has to be annotated and the quality of the performances is closely related to the size of the training corpus. Since the task of annotation is a very tedious one, it is well-worth to find the best way to replace it by an iterative process going as quickly as possible from the semi-automatic steps to the automatic ones.

Taking into account the time dimension, we have attempted to answer the following question: for each particular task, what can be assumed in order to introduce as less as possible noise when automatically enriching the annotated data?

Geraldo Bonorino XexŽo

Programa de Engenharia de Sistemas

Instituto Alberto Luiz Coimbra de P—s Gradua‹o e Pesquisa de Engenharia

Universidade Federal do Rio de Janeiro


Sentiment Recognition in Text Through Affective Computing


Abstract: We present a new paradigm for sentiment analysis in text, Sentiment Recognition, which considers that the externalization of an emotion is subject to noise, due to the author's objectives, target audience and communication venue to transmit the message. The sentiment recognition problem is solved by adapting the noisy channel proposed by Shannon for message communication. To enable this goal two new models are proposed. One is the linguistic emotional model that computes the likelihood of an observed text, given its underlying emotion (hidden). The other one computes the a priori probability of a sentiment in an annotated corpus. The resulting sentiment is the argument of the maximum function, considering three possible polarities (positive,negative and neutral), that multiplies the probabilities from both models. We also show results that demonstrate the efficiency of this new approach.

ClŽment Caron

LIP6, Paris

Workflow Provenance and Data Quality Assessment in WebLab

Abstract: In this talk, we will present a new provenance-based quality assessment
model for data centric workflows. Our goal is to build a tool assisting
workflow designers and users in annotating, exploring and improving the
quality of data produced by complex media mining workflow executions. This is achieved by combining  a rule-based provenance model for generating fine-grained data and service dependencies with a quality assessment model for the controlled propagation of application specific quality annotations through provenance dependency links. We will present the formal semantics of our model and an implementation using standard semantic web technologies within the WebLab media mining platform.


Giseli Rabello Lopes

Departamento de Inform‡tica     

Pontif’cia Universidade Cat—lica do Rio de Janeiro


Approaches to the dataset interlinking recommendation problem


Abstract: Whenever a dataset t is published on the Web of Data, an exploratory search over existing datasets must be performed to identify those datasets that are potential candidates to be interlinked with t. This talk introduces and compares two approaches to address the dataset interlinking recommendation problem, respectively based on Bayesian classifiers and on Social Network Analysis techniques. Both approaches define rank score functions that explore the vocabularies, classes and properties that the datasets use, in addition to the known dataset links. After extensive experiments using real-world datasets, the results show that the rank score functions achieve a mean average precision of around 60%. Intuitively, this means that the exploratory search for datasets to be interlinked with t might be limited to just the top-ranked datasets, reducing the cost of the dataset interlinking process.




Thursday September 11

Mohand Boughanem,

IRIT Toulouse

Social media search: leveraging user generated content


Abstract: Social Web (Web 2.0) technologies has enabled people to express their opinions, to share content (photos, blog posts, videos, bookmarks, etc.); to connect with other users, either directly or via common interests often reflected by shared content; to add free-text tags or keywords to content; users comment on content items. All these user-generated contents UGC need not only to be indexed and searched in effective and scalable ways, but they also provide a huge number of meaningful data, metadata that can be used as clues of evidences in a number of tasks related particularly to information retrieval. Indeed, these user-generated contents have several interesting properties, such as diversity, coverage and popularity that can be used as Òwisdom of crowdsÓ in search process. This presentation will provide some general properties of these data and then briefly lists some search tasks that leverage these data. We will particularly focus on two specific tasks namely microblog search and exploiting UGC to improve  search.

Eric Gaussier and

HervŽ Martin,

LIG Grenoble


General presentation of the Grenoble Informatics Lab. (LIG)

and (b) Content diffusion in social networks.


Abstract: (a) In this presentation, we will describe the scientific project of the LIG, that is "ambient and sustainable IT". The goal is to leverage the complementary nature and recognised quality of the 22 research teams of the LIG to contribute to fundamental aspects of the discipline (modelling, languages, methods, algorithms) and to create a synergy between the conceptual, technological and societal challenges that surround this theme.

(b) Content diffusion : Social networks now play a central role for sharing information and discussing different types of events. The way information spreads in such networks has often been compared to the way innovations spread in marketing or viruses spread in populations. As such, two of the more popular information diffusion models, the IC (Independent Cascade) and the LT (Linear Threshold) models, can be seen as instances of the standard SI (Susceptible-Infectious) family used in epidemiology. However, such models usually fail to account for important characteristics of the users sharing and diffusing information in social network, namely the content of the information being disseminated. After a presentation of the standard information diffusion models, we will introduce a new generation of models which provide a more realistic modeling of how content spreads in social networks.

Jonice Oliveira, UFRJ

Big Social Data: Analyzing and extracting knowledge from social data in web


Abstract: Currently, with the growing of available data   observed as a result of human interactions that are increasingly recorded via web, mobile devices, distributed sensors or revealed through digitization of historical records - we have the opportunity of evaluate and develop techniques for social network analysis. In response to these massive new sources of data, new challenges and opportunities arise from socially-generated big data in several fields (such as science, security, marketing É). The CORES Lab is engaged in some of these challenges. In this talk I will discuss recent results (and some ongoing work) obtained in data extraction and mining. Also, I will describe some works and applications in link recommendation, emergency management, diffusion of information and influence over them.


Jacqueline Leta, UFRJ

Brazilian research in Scientometrics: history, trends and challenges

Abstract: The presentation will focus on three dimensions of Brazilian Scientometrics: history, present main trends and challenges. It will start on the pioneer studies developed at IBICT, the first Brazilian Research Institute to carry on research in Scientometrics. Then, a present picture of Brazilian studies in the field will be shown, highlighting the recent growth of scientometric community in the country as well as its main origin, thematic and type of analysis. Finally, it will be presented some of the main challenges Brazilian researchers in Scientometrics face nowadays, including the difficulties to access/process data of Brazilian output and the lack of expertise in frontier fields, such as Computer sciences

Guillaume Cabanac

IRIT- Univ. Paul Sabatier de Toulouse

Scientometric Studies of Academic Writing


Abstract: The interdisciplinary field of science referring to the quantitative study of science is called "scientometrics."  This talk outlines recent scientometric findings on academic writing.  In collaboration with colleagues in computing, psychology, and the sociology of science, we tackled a variety of questions about gender-specific writing styles, recognition through eponymization, work-life balance in academia, order effects in paper bids for peer-reviewed conferences...

JosŽ Antonio de Macedo          

Departamento de Computa‹o

Centro de Cincias

Universidade Federal do Cear‡


Trajectory Enrichment using Linked Data


Abstract. Mobility analysis is key for several applications domains, such as transportation logistics, traffic management, animal migration, infectious disease spread analysis, to name a few. In the recent years, the availability of movement data has increased incredibly as a consequence of current GPS, GSM, RFID, and sensors technology. However, the effective analysis and understanding of huge amount of mobility data is highly dependent on the application domain semantics. Despite recent works in trajectory data management, there is still a gap between the spatio-temporal aspects and the semantics involved. In this talk we present a process for trajectory semantic enrichment that aiming to reduce this gap. In our approach we benefit from growing collections of linked data already available on the Web in order to add semantics to trajectory data. In addition, we show an architecture that enables integrating sets of trajectories with linked data sources in a systematic way. We finish our talk showing a real case study, where trajectories of vehicles are used to characterise typical behaviour of people travelling in Milan city.

Karine Zeitouni,

PRISM, Univ. Versailles Saint-Quentin


Spatial Big Data Handling - Examples in Astronomy and Transport


Abstract: The amount of data originating from sources like individual phones traces, vehicle and vessel monitoring, environmental sensors, or sky surveys, represent trillions of bytes per day.  These data are mainly geo-referenced, i.e. spatial or spatio-temporal. However, most of the algorithms and techniques used for spatio-temporal management and analysis fail when applied to massive data volumes. This talk intends to introduce the specific problems of handling such Spatial Big Data, and reviews some recent works. To highlight this domain, we will refer to an application example in astronomy related to the MASTODONS project GAIA, and another related to road trajectory data management.

Mohand-Said Hacid, LIRIS Lyon


Aggregated Search in Graph Database: A Relational-Based Approach


Abstract: In this talk, we will discuss the problem of assembling fragments from different graphs to build an answer to a user query. The goal is to be able to provide an answer, by aggregation, when a single graph cannot satisfy all the conditions that are specified in a query. We will discuss a basic algorithm, a relational framework, together with some experiments.

Nicolas Thome

LIP6 Paris

Representation Learning for image understanding


Abstract: The semantic annotation of visual data (images, videos) is one of the most challenging task in computer vision. To achieve this goal, the last decade has witnessed the supremacy of handcrafted powerful image or video local descriptors (e.g. SIFT, STIP). On the other hand, deep learning is appealing due its capacity to learn the features themselves from data, and recently focused attention due to its successful application for large scale image classification (ILSVRC2012-2014). In this talk, I present some of our recent works to learn representations from visual data, using both supervised and unsupervised strategies.

  Firstly, we address the challenging problem of categorizing video sequences composed of dynamic natural scenes. We propose here to represent videos using unsupervised learning of motion features. We show that Slow Feature Analysis (SFA) can be successfully applied to learn the the principal and more stable motion components of training videos. Embedded into a global coding/pooling architecture, SFA provide an effective signature for each video sequence, that is shown to outperform state of the art classification performances on two challenging natural scenes data sets.

  Secondly, we propose a new supervised metric learning scheme. Working with inequality constraints involving quadruplets of images, our approach aims at efficiently modeling similarity from rich or complex semantic label relationships. From these quadruplet-wise constraints, we propose a similarity learning framework relying on a convex optimization scheme. We also propose a new method to regularize the matrix distance learning. We then study how our metric learning scheme can exploit specific class relationships, such as class ranking (relative attributes), and class taxonomy. We also evaluate our approach in a new application to learn similarities between webpage screenshots in a fully unsupervised way.

Renata Mendes de Araujo

Departamento de Inform‡tica Aplicada

Centro de Cincias Exatas e Tecnologia

Universidade Federal do Estado do Rio de Janeiro


Collaborative and Open Government Systems


Abstract: CiberDemocracy is defined as the study on conceptualization, design and use of technology to support and change the dialogue between Citizens and the State, through the creation of innovative processes and online tools for effective discussion, decision making and co-creation between citizens and the public administration. This presentation will describe the main research programs under development at the Research and Innovation Group in CiberDemocracy at the Federal University of State of Rio de Janeiro (UNIRIO). Particularly, it will detail two current joint research initiatives between CiberDem and french universities - UniversitŽ Versailles and UniversitŽ d«Avignon - started in the context of the Brazilian Institute for Web Science Research."

Guillaume Marrel

UniversitŽ d'Avignon et des Pays de Vaucluse


Tr@nspolo : Focus on the analysis of the staging of everyday political work between political and computer sciences


Abstract: The Tr@nspolo project deals with political and organizational transparency in digital societies. It proposes an interdisciplinary discussion between computer and political sciences, on the actors and the uses of participatory Web by exploring two axes. The first axis aims to capture the changes in politics imposed by the request of transparency. The second axis deals with the openness provided by the development of participatory web as a tool for political participation. The presentation will focus on the first axis on the web in political work. It is to explore how digital tools are needed to elected officials and the manner in which they adjust their practices to meet the requirements of transparency and participation.

The research focuses here on one particular type of digital tools: electronic personal organizers considered as tools of management for time of political work, as tools of political communication on the Web, but also as potential instruments of politicians activity control by citizens. The combination of political and computer sciences can contribute to develope scientific tools for identifying data and events, or methods to reconstruct the sequence of events by the overlap of information available on the Web. This is to compare the content of a corpus of data from an electronic calendar (ical) with the media coverage of these events on the Web. This research could lead to the production of a working prototype to make visible and to compare the different dimensions of the work of elected officials (business volume, speakers, territorial coverage, political networks ...). In computing, beyond the resumption of expertise in automatic annotations in terms of named entities and systems of questions/answers, the project has the advantage of addressing a new problem: make a automatic response to a question posed in natural language, which assumes that we know to connect the named entities.

MH Sa Vilas Boas,

Magali Nonjon,

Alexandre Reiffers and

Yonathan Portilla

Agorantic, Univ. avignon


Developing expertise in the design of digital democracy : a comparative and transdisciplinary approach (France-QuŽbec-Brazil)


Abstract: Our presentation will present our on-going research on the developing expertise in the design of digital democracy. Our pluridisciplanary team wants to analyse the engeeniering (actors and tools) engaged in the promotion of digital democracy in France-QuŽbec-Brazil. For this presentation we focused on two main points and one specific area : Brazil.

Firstly, we will focus on a specific experience : the online participatory budgeting. This model is supposed to go further a famous institution which aims at defining local public budget with citizens, the participatory budgeting. The online participatory budgeting has spread around the country and in 2014, there were 37 experiences led by Brazilian cities. A special aim will be to analyse with a multidisciplinary approach the design of tools towards transparency and open dialogue which are used in those experiments.


Secondly,  we will present the results of an on-going work dedicated to the role of scientists in the promotion and the spread of digital tools which are supposed to enhance democracy and citizens' participation. Our question is therefore to understand what part does researchers play in the diffusion of such a model within Brazil. For that purpose, we will use a transdiciplinary approach and we will analyze an original database: the CV lattes. This database is interesting because it gathers the curriculum of most Brazilian researchers. It also gives a very detailed view of their activity, within the academic field or in other ones. We will use web sciences' analysis to reconstitute researchers' communities around the online participatory budgeting: who and where are the researchers who study such a tool ? Is there a nodal group within these communities, which play a central of diffusion of the online participatory budgeting? Moreover, for which institutions or agency these researchers produce analysis ?

Yahia Taher, PRISM Univ. Versailles Saint Quentin


Citizen-Centric Business Process


Abstract: In this presentation we provide a description of an approach to specify co-participation requirements in the domain of public service delivery in order to allow flexibility in the high-level requirements identification for citizen-government supporting tools according to targeted co-participation (citizen – public administration) levels. It discusses how fine-tuned collaboration support, public service information transparency and social memory can be designed over a service-oriented view for public service design and delivery to progressively enhance government-citizen interaction levels.



Friday September 12

Esther Pacitti, INRIA&Lirmm, UM2

Profile Diversity for Query Processing using Users Recommendations

Abstract: Many scientific fields produce and consume a considerable amount of diverse data (e.g. biology , astronomy, physics) stored in different heterogeneous sites, and produced by different types of users profiles. We investigate two different use cases: a) In the domain of plant phenotyping, there has recently been increasing interests in finding diverse data coming from different research communities. b)In botany, the emergence of citizen sciences has fostered the creation of large and structured communities of nature observers. In this context, there is a need to retrieve diverse plant observations from a diverse spectrum of plant families, genus and species. In this talk I will present some new issues of profile diversity, a novel idea in searching and recommending scientific items (e.g. documents, images, datasets, etc), and how profile diversity can be deployed in different kinds of infrastructures (centralized and distributed).

Sean Wolfgand Matsui Siqueira Departamento de Inform‡tica Aplicada

Centro de Cincias Exatas e Tecnologia

Universidade Federal do Estado do Rio de Janeiro


Analyzing Social Presence and Willingness in Online Social Networks


Abstract: Online communities have become important places for users to exchange information and build knowledge. In these communities, the social presence shows how a participant projects social and emotionally its presence while it is also related to his willingness to participate and collaborate with the others. We have developed some works to capture the social presence in online communities, specially with the aim of supporting tutors to guide learners in building knowledge. In addition, we have proposed some approaches to identify people who are willing to help and provide good answers in online communities of questions and answers.

Florence Andreacola,

FR Agorantic, LIA, UAPV


Museum, web and equipped visitors. Interactions and appropriation of knowledge by participatory tools


Abstract: During the visit of a museum exhibition and through this activity, relationships are established between visitors themselves and between visitors and museum institution. The complex object that is an exhibition put visitors in touch with meaningful objects, scenography, texts, knowledge, valueÉ To these many-facetted situation proposed by the curator and interpreted by visitors, we have to consider the relations that visitors build together during the visit too. So the question of the impact of digital devices (and the access to the Internet they allow) on the visitÕs activity has emerged in the context of the increase of these devicesÕ use. During this communication, we would draw up a state of the research that has for object the study of the interactions visitors engaged with other visitors and with the museum they visit. In this context, we would observe and analyse the potential impacts of digital technologies and the web on this relationship established around and about the social and symbolic object that is museum exhibition.

Engelbert Mephu Nguifo

Limos, Univ. Clermont Ferrand


Towards more targeted recommendations in folksonomies


Abstract: Recommender systems are now popular both commercially as well as within the research community, where many approaches have been suggested for providing recommendations. FolksonomiesÕ users are sharing items (e.g., movies, books, bookmarks, etc.) by annotating them with freely chosen tags. Within the Web 2.0 age, users become the core of the system since they are both the contributors and the creators of the information. It is of paramount importance to intercept their needs for providing a more targeted recommendation. In our talk, we will introduce a new dimension in a folksonomy classically composed of three dimensions <users,tags,resources> and propose an approach to group users with close interests through quadratic concepts. Such structures are used in order to propose our personalized recommendation system of users, tags and resources. We will finally summarize extensive experiments that we conduct on two real-life datasets (MovieLens and BookCrossing) in terms of precision and recall as well as social evaluation.


Antonia Alcaraz

CNRS-ChargŽe de programme AmŽrique latine


Cooperation tools between France (Europe) and Brazil (South America)


Abstract: Brazil is the main scientific partner of the CNRS in Latin America. Institutional cooperation with Brazil, in particular with CNPq referred to more than 40 years ago has given high quality international results. This cooperation aims at structuring scientific collaboration and taking advantage of the regional synergies. The presentation will be in four parts : 1) CNRS cooperation tools : International Research Networks (GDRI) are research networks which associate many laboratories from several countries on a specified scientific theme. 2) CNRS Cooperation with Brazil : focus on GDRI Websciences. 3) French, European and Brazilian additional support. 4) Key dates and contacts.