Group seminars

In this talk, I will first provide a quick overview of our works in the area of graph query processing. Then I will provide some details on our works on answering reachability and historical reachability queries. A reachability query asks, given two vertices u and v in a directed graph G, whether u can reach v, that is, whether there is a path from u to v in G. This problem has been well studied and there is a large body of literature. We show how a multi-level compression technique can speed-up reachability query answering.

In this talk we are going to discuss our experience in building a real ICT Ecosystem to support the data life cycle for Precision Medicine Projects and provide our thoughts on how to use knowledge graphs methods for managing the data and produced knowledge. The systems are used by clinicians to record clinical and genomic data and support a fully digitized data life cycle.

As informavores, information seeking is a key characteristic of human nature. Fueled by curiosity, humans usually navigate a plethora of real-world networks, including but not limited to the World Wide Web, online encyclopedic systems, news articles, and social networks. Consequently, the navigation patterns employed by humans provide deeper insights into how humans explore, browse, and interface with information on the Web.

An emerging trend is to leverage an under-explored and richly heterogeneous pool of human knowledge inside machine algorithms, a practice popularly termed as human-in-the-loop (HIL) process. A wide variety of applications, starting from query processing to text translation, feature engineering, or even human decision making in complex uncertain environments stand to benefit from such synergistic man-machine collaboration. This talk will give an overview of our ongoing research projects, recent research results, and impacts.

The evolution of ICT is a rapid and inexorable process that changed the society landscape several times in the last decades.

Exploration is one of the primordial ways to accrue knowledge about the world and its nature. As we accumulate, mostly automatically, data at unprecedented volumes and speed, our datasets have become complex and hard to understand. In this context, exploratory methods offer the capabilities of progressively gathering the necessary knowledge when dealing with datasets that are to us "terra incognita". Yet, when dealing with complex data, we are also in need of powerful data models that give us the necessary expressivity to properly handle the richness and intricacies of the data at hand.

A key step in the democratization of analytics is the democratization of Extract Load Transform (ELT). Democratization depends on SQL as SQL is the most well-known, declarative language. Indeed, SQL adoption increases in areas that were not considered the traditional domain of SQL; a new class of users, often called analytics engineers, are enabled to SQL their way into areas that were previously inaccessible to them.

Data is becoming more and more accessible. The proliferation of Web platforms (e.g., social media and web fora) together with the increased affordance of smart sensing technologies allowed accessing data about a large and diverse set of users, as their activities in the digital world reflect their real-life desires, drives, and needs. In the last decade(s), several computational solutions - based on data science and machine learning -- have been developed to make sense and extract useful and meaningful insights from this large amount of data.

Datalog and its extensions (such as existential rules) are widely studied querying formalism. A natural question is to characterise which queries can be expressed within such query languages, in the style of Immerman-Vardi theorem. In this talk, I will present some results regarding the expressivity of Datalog-based query languages. The last such result if that, perhaps surprisingly, any query that is decidable and closed under homomorphism is expressible by a set of existential rules for which the standard chase is universally terminating.

Temporal property graphs are graphs whose structure and properties change over time. Temporal graph datasets tend to be

For several decades the study of graph query languages has concentrated on RPQs (regular path queries) and many derivatives: CRPQ, 2RPQ, 2CRPQ, ECRPQ, RDPQ, to name a few. But do these really constitute the core of graph query languages? A quick look at Cypher and the emerging GQL and SQL/PGQ standards lead to the negative answer. But what shall we be studying then, how, and why? This talk briefly surveys the classics, outlines key features of graph languages, and suggests new formal models as well as ways to influence the new standards.

Data streaming enables online monitoring of large and continuous event streams in Cyber-Physical Systems (CPSs). In such scenarios, fine-grained backward provenance tools can connect streaming query results to the source data producing them, allowing analysts to study the dependency/causality of CPS events. While CPS monitoring commonly produces many events, backward provenance does not help prioritize event inspection since it does not specify if an event's provenance could still contribute to future results.

Top-k queries have been studied intensively in the database community and they are an important means to reduce query cost when only the “best” or “most interesting” results are needed instead of the full output. However, optimality results for queries involving joins hold only in a fairly limited model of computation that does not account for the cost incurred by large intermediate results.

Node classification is an important problem in graph data management. It is commonly solved by various label propagation methods that work iteratively starting from a few labeled seed nodes. For graphs with arbitrary compatibilities between classes, these methods crucially depend on knowing the compatibility matrix that must be provided by either domain experts or heuristics. Can we instead directly estimate the correct compatibilities from a sparsely labeled graph in a principled and scalable way?

The distributed nature of Web applications is one of the reasons for the popularity of the Web.

There is an increasingly pressing need, by several applications in diverse domains, for developing techniques able to analyze very large collections of sequences, or data series. Examples of such applications come from astrophysics and a multitude of other scientific and application domains that need to apply machine learning techniques for knowledge extraction. It is not unusual for these applications to involve numbers of data series in the order of hundreds of millions to billions, which are often times not analyzed in their full detail due to their sheer size.

The Software Heritage project has assembled the largest existing archive of publicly available software source code and associated development history, for more than 6 billion unique source code files and 1 billion unique commits, coming from more than 90 million software development projects.

Pattern matching in graphs, that is, finding subgraphs that match a smaller template graph within a large background graph, is fundamental to graph analysis and serves a rich set of applications. Unfortunately, existing solutions have limited scalability, are difficult to parallelize, and/or support only a limited set of search patterns.

We explore avenues towards a scalable solution for pattern matching. We target practical pattern matching scenarios in large-scale property graphs, able to operate on massive distributed memory machines.

(Titre Fr. Fouille de Modèles Exceptionnels dans les Données Comportementales)

We strive to find contexts (i.e., subgroups of entities) under which exceptional (dis-)agreement occurs among a group of individuals, in any type of data featuring individuals (e.g., parliamentarians, customers) performing observable actions (e.g., votes, ratings) on entities (e.g., legislative procedures, movies). To this end, we introduce the problem of discovering statistically significant exceptional contextual intra-group agreement patterns. To handle the sparsity inherent to voting and rating data, we use Krippendorff’s Alpha measure for assessing the agreement among individuals.

Data scientists spend big chunk of their time preparing, cleaning, and transforming raw data before getting the chance to feed this data to their well-crafted models. Despite the efforts to build robust predication and classification models, data errors still the main reason for having low quality results. This massive labor-intensive exercises to clean data remain the main impediment to automatic end-to-end AI pipeline for data science.

Les technologies du Web Sémantique offrent de nouvelles possibilités pour améliorer les services dédiés à la diffusion des connaissances culturelles et intellectuelles. Cependant, de nombreuses institutions, en charge de fonds documentaires, doivent gérer leurs catalogues selon des normes héritées des années 60. Le paradigme historique des notices papier, où chaque notice décrit un document possédé par une institution, est encore majoritairement utilisé par la communauté documentaire.

Les technologies du Web Sémantique offrent de nouvelles possibilités pour améliorer les services dédiés à la diffusion des connaissances culturelles et intellectuelles. Cependant, de nombreuses institutions, en charge de fonds documentaires, doivent gérer leurs catalogues selon des normes héritées des années 60. Le paradigme historique des notices papier, où chaque notice décrit un document possédé par une institution, est encore majoritairement utilisé par la communauté documentaire.

Retour sur le travail fait sur la plateforme ExpeData:

• expérimentation HelloWorld et bonnes pratiques expérimentales
• Stream Management System

Le traitement de flux de données est au cœur des problématiques actuelles liées au Big Data. Face à de grandes quantités de données (Volume) accessibles de manière éphémère (Vélocité), des solutions spécifiques tels que les systèmes de gestion de flux de données (SGFD) ont été développés. Ces SGFD reçoivent des flux et des requêtes continues pour générer de nouveaux résultats aussi longtemps que des données arrivent en entrée.

While schema mapping specification is a cumbersome task for data curation specialists, it becomes unfeasible for non-expert users, who are unacquainted with the semantics and languages of the involved transformations.

When using machine learning algorithms to solve classification problems, one recurring problem is the one of unbalanced datasets, especially in binary classification when one class is much bigger than the other. Several solutions have been proposed to tackle the problem, such as undersampling of the majority class. This undersampling approach is generally based on statistical on the data. We propose a different approach to this problem, by considering the functional dependencies and their interactions in the two classes.

In today’s area of Big Data, data is collected to a large extent automatically by hard- and software sensors in fine granularity and low abstraction. Where users interact with data, e.g. for data analytics, they typically think, reason, and talk about entities of larger granularity and higher abstraction. For instance, a graph of twitter communication contains individual messages, retweet relationships, etc. while social network analysis done on such data is interested in discussions, topics, communities and so on.

1. Réflexions sur la valorisation de la production logicielle Liris-BD
2. Poursuite des réflexions sur la reproductibilité d'expérimentations

Within the big data tsunami, relational databases and SQL are still there and remain mandatory in most of cases for accessing data. On the one hand, SQL is easy-to-use by non specialists and allows to identify pertinent initial data at the very beginning of the data exploration process. On the other hand, it is not always so easy to formulate SQL queries: nowadays, it is more and more frequent to have several databases available for one application domain, some of them with hundreds of tables and/or attributes.

We address the problem of discovering contexts that lead well-distinguished collections of individuals to change their pairwise agreement w.r.t. to their usual one. For instance, in the European parliament, while in overall, a strong disagreement is witnessed between deputies of the far-right French party Front National and deputies of the left party Front de Gauche, a strong agreement is observed between these deputies in votes related to the thematic: External relations with the union.

(répétition avant soutenance de thèse)

De nos jours, on a de plus en plus de capteurs qui ont tendance à apporter confort et facilité dans notre vie quotidienne. Ces capteurs sont faciles à déployer et à intégrer dans une variété d’applications (monitoring de bâtiments intelligents, aide à la personne,...). Ces milliers (voire millions)de capteurs sont de plus en plus envahissants et génèrent sans arrêt des masses énormes de données qu’on doit stocker et gérer pour le bon fonctionnement des applications qui en dépendent.

(répétition avant soutenance de thèse)

Les systèmes d’aide à la décision occupent une place prépondérante au sein des entreprises et des grandes organisations, pour permettre des analyses dédiées à la prise de décisions. Avec l’avènement du big data, le volume des données d’analyses atteint des tailles critiques, défiant les approches classiques d’entreposage de données, dont les solutions actuelles reposent principalement sur des bases de données R-OLAP.

Semantic interoperability is said to be the enabler for 40-60% of the global economic impact of the Internet of Things at the horizon 2025. In this talk I propose an overview of some of the main scientific problems that need to be solved to reach this goal, along with some of my research and standardization contributions. 1. how to make the Things able to securely interact on the Web in conformance with its architecture principles, and the REST architectural design principles?

Recent researches and analysis reports declare that high energy consumption of buildings is major problem in developed countries. As a result, they show concretely that building energy management systems (BEMS) and deployed wireless sensor network environments are important for energy efficiency of building operations. In the literature, existing smart building management systems focus on energy consumption of the building, hardware deployed inside/outside of the building and network communication issues.

Répétition avant soutenance de thèse : Cette thèse porte sur la mise en place d'un système de requêtage sur des flux sous contraintes de capacités. Ce système est porté par ses utilisateurs⋅trices et basé sur les similitudes entre requêtes. Les relations d'équivalences entre les différentes requêtes permettent de réunir les participants au sein de communautés d'intérêt. Celles-ci forment alors une abstraction permettant de séparer le problème d'organisation du système en plusieurs sous-problèmes plus simples et de taille réduite.

In many applications, such as social networks, production lines or stock analysis, its is quite essential to create and process large amount of data at high rates. Such continuous and unbounded flow of data is abstracted under the concept of data streams, and data items within such streams are generally called as events. Traditional Data stream processing techniques -- that are based on the relational models -- deal with a wide variety of applications, however, heterogeneity among the data sources makes it difficult to provide a generalised solution.

One of the core visions of the Semantic Web is that data can be shared across the boundaries of applications and websites. A particular application of Semantic Web is Linked Data, which publishes data as Web Data in RDF format. It enables sharing and accessing data in a decentralised manner. This is has led to the rise of many Linked Data initiatives, with its most successful project Linked Open Data (LOD), which at this time numbers thousands of datasets, and with a magnitude of billions of triples.

Massive graph data sets are pervasive in contemporary application domains. Hence, graph database systems are becoming increasingly important. In the study of these systems, it is vital that the research community has shared benchmarking solutions for the generation of database instances and query workloads having predictable and controllable properties. Similarly to TPC benchmarks for relational databases, benchmarks for graph databases have been important drivers for the Semantic Web and graph data management communities.

Ubiquitous computing and modern data collection tools have given rise to a new era of data management, in which the data as well as the processing environments are enormous and diverse. In this era in which the terms 'Big Data' and 'Cloud Computing' prevail, the success of data management relies in optimizing data processing on multiple dimensions and delivering data services with multifarious quality guarantees. This talk discusses the challenges and the opportunities of the new era of data management and explores the possibilities and limitations of multi-dimensional data management.

In a large number of domains, ranging from astrophysics to earth observation, data analysts are facing a data deluge. In this Big Data era, it is essential to explore the data in order to unearth new knowledge. As user profiles are becoming more and more diverse and data ever more complex, this task has become increasingly hard. Analysts can access gigantic scientific data through SQL, while also using data mining tools to peer their data. We propose a query rewriting technique to help data analysts formulate their queries, in order to rapidly and intuitively explore their Big Data.

Dans le cadre d'une candidature au poste MC ECL-LIRIS : Je présenterai rapidement mes travaux de thèse sur la sécurité des implémentations cryptographiques face aux attaques physiques, puis je parlerai de ce que le fais en post-doc autour de la notion de "privacy as control". Je finirai par donner un aperçu de ce que j'aimerais faire par la suite dans le cadre de la protection des données personnelles, pour ouvrir la discussion et voir si des intérêts sur ce sujet convergent avec les membres de l'équipe.

Data mining, particularly pattern mining and cluster mining, aims to find interesting and characteristic local structures in data with enumerational approaches, not to miss important ones. However, a big difficulty is on the huge number of solutions so that we can not identify which solutions are really important. This difficulty is a great barrier on the use of pattern mining and cluster mining in practice. In this talk, we propose a new approach called "data polishing" that takes an approach totally different from existing approaches.

Répétition avant soutenance de thèse

Peer-to-peer publish/subscribe architectures are an interesting support for scalable distributed data stream applications. Most approaches, often based on brokers, have a static organization which is not much adaptive to different configurations of the participants’ capacities. We present QTor (Query Torrent) a generic organization that enables dynamic adaptation providing a continuum from centralized to fully decentralized solutions.

In many domains, scientific discoveries rely increasingly on our ability to exploit ever growing volumes of data. A key point is managing the complexity of data life cycles, i.e. the various operations that happen to data from their creation to their deletion: transfer, archival, replication, disposal, etc. These formerly straightforward operations become intractable when data volume grows dramatically, because of the heterogeneity of data management software on the one hand, and the complexity of the infrastructures

Parameter analysis and future disease prediction of health related Big Data are still in an informative stage due to the diversified bulky health care data, which is generated with greater speed. In this talk, data collection architecture and parameter analysis methods of healthcare Big Data will be discussed. Real implementation of ECG batch data in single cluster Apache Hadoop through the MapReduce framework will be presented.

Short Bio:

Journée de présoutenances des master recherche, organisée par le pôle Data Science.

TBA

http://satin-ppl.telecom-st-etienne.fr/flaforest/

Etant donné l’évolution de la notion de texte et la croissance continue des informations textuelles et numériques, de multiples natures, qui sont disponibles en ligne, un des challenges importants pour les linguistes et les analystes de l’information, afin de pouvoir formuler des hypothèses et de valider des modèles, est d’exploiter des outils efficaces pour l’analyse textuelle, capables de s’adapter à des volumes importants de données hétérogènes, changeantes et souvent et de nature distribuée.

In the "big data" era, data integration is a popular activity both in academia and in industry. Integrating hundreds of heterogeneous sources on a daily basis requires a great amount of manual work in order to have data that is polished enough to be useful in the final applications, such as querying and mining. The problem is even harder in practice, as data is often dirty in nature because of typos, duplicates, and so on, that can lead to poor results in the analytic tasks.

This paper proposes a flexible control framework for relational personal data that enforces data originators' dissemination policies. Inspired by the sticky policy paradigm and mandatory access control, dissemination policies are linked with atomic data and are combined when different pieces of data are merged.

CEDAR (Constraint Event Driven Automated Reasoning) is two-year ANR Chair of Excellence project at the LIRIS that has just recently reached completion as per its planned schedule (Jan. 2013 - Jan. 2015). In this presentation, I will overview the issues and challenges taken on by the project - essentially the systematic experimental evaluation of Semantic Web reasoning systems (Track 1) and the management of Big Linked Data (Track 2), as well as proposing our own systems for each track.

The tremendous increase of data sources containing spatial information is bound up with the diversity of geospatial applications such as location-based services (LBS) and global positioning systems. LBS providers use maps to locate spatial entities referring to points of interest (POI), for instance restaurants or locations of events. In our study, we specifically focus on places that tourists can get through LBS.

Présentation dans le cadre du Groupe de Travail "Traitement distribué de requêtes"

RQL (pour Rule Query Language) est un langage de requêtes "à la SQL" qui étend et généralise les dépendances fonctionnelles à de nouvelles catégories de règles. RQL apporte aux analystes de données un outil pratique pour découvrir les implications logiques entre attributs d’une base de données. Ces implications peuvent mettre en évidence des problèmes de qualité de données ou de nouvelles corrélations inattendues entre les attributs. Le traitement de ces requêtes RQL est basé sur une technique de réécriture qui délègue un maximum de calculs au SGBD sous-jacent.

Plusieurs fournisseurs de services de geolocalisation (GoogleMaps, BingMaps, ...) offrent des information à propos des Point d'Intérêts (POIs) et qui les représentent sur des cartes. Les données de ces fournisseurs sont parfois incompletes et parfois inconsistantes entre eux. On propose d'intégrer les POIs de ces fournisseurs pour offrir des données plus précises aux utilisateurs.

9h30:10h10 - (GRAMA) Ghizlane Echbarthi (M2R, Juin): Graphes pour les grandes masses de données en flux [Hamamache]
10h10:10h50 - (BD) Kamel Taouche (M2R TI, Juin): Optimisation d'un algorithme de traitement de requêtes agrégées et approximatives (Suite ANR AOC) [Fabien (Duchateau), Nicolas, Saïd]
10h50:11h30 - (BD) Mehdi Benaïssa (M2R IADE, Juin): Alignement d'entités sémantiques basé sur Map Reduce et sur des sources de référence (PHC KOGAR) [Fabien (Duchateau), Nicolas]

10h00:10h40 - (BD) Ameni Ayadi (M2R, Juin): Exploration des données à base de contre-exemples [Manu, Jean-Marc]
10h40:11h20 - (BD&DM2L) Roland Kotto Kombi (M2R): projet transverse "olfamining" [Fabien De Marchi, Mehdi Kaytoue, Marc Plantevit]

Les bâtiments intelligents sont des bâtiments dans lesquels les nouvelles technologies ont été introduites dans le but de simplifier la vie de l’occupant, d’assurer son confort, de veiller à sa santé, à sa sécurité ainsi qu’à celle du bâtiment. Dans ce cadre, une attention particulière doit être portée à l'économie d'énergie, afin de préserver au mieux l’autonomie des capteurs/actionneurs présents dans le bâtiment et d’assurer la continuité du service.

With the amount of data produced in several application domains, it is increasingly difficult to manage and query related large data repositories. Within the PetaSky project (http://com.isima.fr/Petasky), we focus on the problem of managing scientific data in the field of cosmology. The data we consider are those of the LSST project (http://www.lsst.org/). The overall expected size of the database that will be produced will exceed 60 PB.

Tarek Sayah, doctorant 1ere année dans l'équipe BD, présentera un travail de synthèse sur le contrôle d'accès aux données RDF. Des critères de comparaison sont identifiés (expressivité du langage de sélection des objets, méthodes de résolution des conflits de décisions, ...) et un ensemble de propositions comparées vis-à-vis des ces critères. Un ensemble de directions de travail de thèse sera présenté en conclusion.

Current applications, from complex sensor systems (e.g. quantified self) to online e-markets acquire vast quantities of personal information which usually ends-up on central servers. Decentralized architectures, devised to help individuals keep full control of their data, hinder global treatments and queries, impeding the developmentof services of great interest.

http://pages.saclay.inria.fr/ioana.manolescu/

The two main challenges for the coming to pass of the Semantic Web are scalability and distribution. The problem of scalability is that a well-designed web-oriented knowledge-based system must be able to handle larger and larger volumes of knowledge without unbearable degradation of performance. Dealing with the second challenge — distribution — is as complex an issue since it must deal efficiently and seamlessly with knowledge spread all over the Internet under “real-life” conditions (cache management, handling faulty connections and time delays, query distribution, etc.).

La recherche scientifique et les projets en informatique peuvent être utilisés pour résoudre des problèmes de notre société. Le laboratoire LIRIS développe ce type de recherche dans le groupe de bases de données parmi lesquels il a des projets développés avec le Mexique. La présentation montrera les efforts que nous avons réalisés dans le cadre de notre collaboration. La recherche s’est dirigée vers la génération et la gestion de la connaissance dans le domaine des bases de données géo-spatiales.

Pattern mining over data streams is critical to a variety of applications such as prediction and evolution of weather phenomena or anomaly detection in security applications. Most of the current techniques attempt to discover associations between events appearing on the same data stream but are not able to discover associations over multiple heterogeneous data streams. In this work, we aim to identify temporal dependencies between data streams. We represent event streams by state streams that are induced by the streams' events themselves.

In the market there are several providers for Location-Based Services, and each of them uses its own concepts, models and data. As a consequence, customers obtain different answers from one provider or from another . In this context, the UNIMAP project aims at providing Location-Based Services in which data from various suppliers are aggregated. This ensures the completeness and the correction of the results for a given query. However, the main research question is to deal with the detection and the integration of homologous objects from the databases of these suppliers.

Séminaire mensuel du LIRIS par Sergei O. Kuznetsov, National Research University Higher School of Economics (Moscow, Russia)

http://liris.cnrs.fr/evenements/OnUbiquityOfImplications

Dans le cadre de ma thèse Cifre (collaboration entre l'équipe BD et l'entreprise Systar), je dois travailler sur l'amélioration des performances d'un SGBD bien particulier qui est bitemporel et orienté colonne. Ce séminaire sera l'occasion de présenter le contexte de mes travaux, puis d'expliquer ces deux concepts et enfin de présenter certaines pistes de travail.

Soutenance de thèse blanche.

Voir http://liris.cnrs.fr/membres/?idn=hfeng&onglet=these

http://www.i.kyushu-u.ac.jp/~suzuki/shao-e.html

Business services are increasingly dependent upon Web applications. Whereas URL-based access control is one of the most prominent and pervasive security mechanism in use, failure to restrict URL accesses is still a major security risk.We argue that this risk can be mitigated by providing formal analysis tools to evaluate access control policies as well as the impact of changes on configurations.

Le premier ouvrage de référence de la théorie des treillis est la première édition du livre de Birkhoff en 1940. Cependant, la notion de treillis a été introduite dès la fin du 19ème siècle comme une structure algébrique munie de deux opérateurs appelés borne inférieure et borne supérieure.

Présentation d'un article accepté à http://www.sec2012.org/ suivie d'une discussion sur la continuité du travail.

Information is more and more stored over distributed sources. A simple access to information in these sources requires a single access point. Data integration methods are designed to provide this kind of access by allowing to specify a mediator between the users and the sources. From an access control point of view, the question is how to specify access control of a system built on top of distributed data sources. Each source specifies and enforces its own policies. So, an access control enforced at the mediator level has to preserve the local access controls (of the sources).

Forming multidisciplinary teams is a key to carry out complex tasks, which is increasingly the case higher up in the knowledge value chain. Need for team recommendation systems has always been there, both in product companies or academy. Many studies show that in academic world number of authors per paper and coauthors per author are high and increasing.

Today we are witnessing the commencement of a new information era, the Internet of Things, a planetary web, which interconnects not only data and people, but also quiescent, assorted objects, distributed over large areas. By activating the universe of things surrounding us, we are heading towards Mark Weiser's envisioned world, where computers vanish into the background, fitting smoothly into the user environment and gracefully providing information and services to users, rather than forcing them to adapt to the intricate ambiance from the computing realm.

Service composition has risen from the need to make information systems more flexible and open. The Service Oriented Architecture has become the reference architecture model for applications carried by the impetus of Internet (Web). In fact, information systems are able to expose interfaces through the Web which has increased the number of available Web services. On the other hand, with the emergence of the Web 2.0, service composition has evolved toward web users with limited technical skills.

• WHY SO ? (Theoretical result for LTL-based compliance Based management)
• UMC/SAT-based Solver for Model Checking Business Artifact and Extract Compliance Based Fragments
• Handling Conflicts in Depth-First Search for LTL Tableau to Debug Compliance Based Languages

La gestion des flux de données a connu un grand essor au cours des 10 dernières années. Basée sur le modèle relationnel, elle tente de traiter les flux de données de manière déclarative. Toutefois, à cause de sa grande complexité, pour la même requête, les systèmes existants produisent des résultats différents. Astral est une algèbre de gestion de flux de donnée qui permet de clarifier les sémantiques des opérateurs d’un tel système. Afin de montrer que cette algèbre est utilisable, un prototype a été conçu pour traiter les flux tels que décrits par l'algèbre.

Nowadays, many enterprises and organizations are faced to large volumes of data that have to be analyzed in a per-day basis. In particular, scientific datasets are growing at unprecedented rates and are likely to continue growing to the order of Exabytes. These current needs of data management require applications to run over a large number of computing nodes. However, databases management systems (DBMS) have proven inefficient to deal with very large datasets as well as to scale out to a large number of computing nodes.

Peer to peer systems have been widely used to alleviate the burden of servers by transferring to peers in a network tasks that do not require a centralization of the information. A wide range of applications are now emerging over peer-to-peer, such as social networking, multiplayer games, mobile messaging, etc.

Most of these applications are essentially data-centric, they rely on exchange of data between peers, and can be expressed by queries over the database.

keywords: Email, Information Extraction, Semantic query, Text mining.

En collaboration avec Pierre-Antoine Champin de l'équipe SILEX, nous avons mis en place un serveur fournissant les données publiques du LIRIS en RDF. Tandis que ces données étaient auparavant exposées au sein de pages Web HTML, elles sont maintenant accessibles soit directement comme des fichiers indépendant en RDF ou par le biais d'un point d'accès SPARQL. Le serveur en question génère les données directement depuis la base de données relationnelle du LIRIS. Antoine présentera l'architecture et fera une démonstration du système qui sera rendu accessible au monde très prochainement.

Cet exposé intitulé qui fait suite aux travaux effectués dans le cadre du projet OPTIMACS. Résumé: La multiplication d'applications reposant sur des réseaux de capteurs distribués engendre un intérêt croissant pour la modélisation d'applications "pervasives" dans des environnements dynamiques. Ce type d'applications touche des domaines tels que la prévention de risques de tremblements de terre ou d'inondations, la traçabilité de personnes ou d'objets, etc.

La gestion de données de référence, plus connue sous le sigle MDM pour « Master Data Management », est la discipline des technologies de l'information qui s'applique aux données de référence partagées par plusieurs Systèmes d’Information (SI), systèmes informatiques ou bases de données...

In this talk, we present some of our works, biased toward our side, done under a Japanese-French project on applying machine learning and data mining to the design problem of swarm robot controllers. Recently we have witnessed a remarkable progress both in hardware and software concerning mobile robots that learn and discover. Our applications concern several tasks including patrol, formation control, human avoidance, and discovery, which are accomplished in simulator, closed arena, and office environments.

Peers in a peer-to-peer data management system often have heterogeneous schemas and no mediated global schema. To translate queries across peers, we assume each peer provides correspondences between its schema and a small number of other peer schemas. We focus on query reformulation in the presence of heterogeneous XML schemas, including data-metadata conflicts. We develop an algorithm for inferring precise mapping rules from informal schema correspondences. We define the semantics of query answering in this setting and develop query translation algorithm.

L’abondance des motifs générés par les algorithmes d’extraction de connaissances représente un grand problème dans le domaine de fouille de données. Afin de faciliter l’exploration de ces motifs, deux approches sont souvent utilisées : la première consiste à résumer les ensembles de motifs extraits et la seconde approche repose sur la construction de représentations visuelles de ces motifs. Cependant, les résumés ne sont pas structurés et ils sont proposés sans méthode d’exploration.

Je donnerai un aperçu de haut niveau des concepts du Web des données, du Web sémantique, des principes des “Linked Data” ainsi que des ontologies du Web en montrant que, tout en favorisant l'interopérabilité, la réutilisabilité et l'échange de données, le développement et le déploiement de ces technologies à large échelle sont encore truffés de pièges et semés d'embuches théoriques et pratiques. J'insisterai tout de même sur les succès – relatifs – accomplis dans ce domaine.

Les nouvelles applications rendues possibles par internet et les nouveaux besoins qu’elles engendrent font exploser les volumes d’informations. La prise en charge de cet accroissement très rapide est d’autant plus complexe que les données sont très diverses. Cette diversité est due tant à la nature des données elles-mêmes (texte, multimédia, méta-données, traces) qu’à la manière dont elles sont produites (senseur, personne, groupe de personnes…).

Les organisations doivent se conformer à de nombreuses règles de droit issues de la loi, de règlements ou d'engagements contractuels.
Toutes ces règles partagent des traits caractéristiques comme la présence de déclencheurs qui les activent, d'échéance où elles sont évaluables ou de sanctions/réparations associées.
Un problème d'intérêt est de vérifier la conformité de ses règles sur des traces issues de l'activité des organisations, dans une perspective d'audit.
Nous avons conçu un langage formel dédié à ces règles, doté d'une sémantique dénotationnelle.

In this talk, we present two powerful, fully automated methods to
evaluate Datalog programs: the first approach transforms the Datalog
program in an implicit Boolean Equation Systems (BESs) solved by
existing general purpose verification toolboxes, such as CADP,
providing local BES resolutions with linear-time complexity; the
second approach transforms Datalog programs into Rewriting Logic and
produces an efficient rewrite system exploiting the main features of
the high-level programming language Maude. We confirm our results

L'informatique ambiante est une réalité qui s'impose amenant avec elle des défis majeurs qui touchent à divers domaines comme les intergiciels, les bases de données et les réseaux de communication pour ne citer que ceux là. Une des problématiques majeures de l'informatique ambiante est comment utiliser et surtout composer les services offerts par cette dernière, de manière à répondre aux attentes des usagers. Dans cet exposé, je vais définir et formaliser une composition dynamique des services qui touchent à leur aspects fonctionnels, non fonctionnels et structurels.

13:00 Brice Chardin : Historisation de données et impact des mémoires flash
13:30 Rosa Marina Donolo : Indicateurs spatiaux environnementaux à échelle urbaine : définition, implémentation et visualisation
14:00 François Hantry : High level Business queries
14:30 .... pause
14:45 Roula Karam : Intégration et Visualisation des services LBS provenant de plusieurs fournisseurs
15:15 Marius Massala : Métadonnées géographiques et recherche d'information

Notre but est d'analyser une approche pour définir et découvrir des mappings probabilistes entre deux taxonomies. D'abord, nous comparons 2 façons de modéliser les mappings probabilistes qui sont compatibles avec les contraintes logiques déclarées dans les taxonomies. Nous suivons une démarche bayésienne pour estimer les probabilités des mappings, en exploitant les descriptions des instances associées aux classes des taxonomies.