Information available online is more and more complex, distributed, heterogeneous, replicated, and changing. Web services, such as SOAP services, should also be viewed as information to be exploited. The goal of Gemo is to study fundamental problems that are raised by modern information and knowledge management systems, and propose novel solutions to solve these problems. A main theme is the integration of information, seen as a general concept, including the discovery of meaningful information sources or services, the understanding of their content or goal, their integration and the monitoring of their evolution over time.
Gemo works on environments that are both powerful and flexible to simplify the development and deployment of applications providing fast access to meaningful data. In particular, content warehouses and mediators offering a wide access to multiple heterogeneous sources provide a good means of achieving these goals.
Gemo is a project born from the merging of INRIA-Rocquencourt project Verso, with members of the IASI group of LRI. It is located in Orsay-Saclay. A particularity of the group is to address data and knowledge management issues by combining techniques coming from artificial intelligence (such as classification) and databases (such as indexing).
Databases do not have specific application fields. As a matter of fact, most human activities lead today to some form of data management. In particular, all applications involving the processing of large amounts of data require the use of databases. Technologies recently developed within the group focus on novel applications in the context of the Web, telecom, multimedia, enterprise portals, or information systems. They often consider data integration as in the application of food risk that we mention next.
The warehouse allows to acquire information from the Web about food risk, to enrich this information, e.g., by classifying and integrating it, and finally provide a unique entry point for it.
Our goal is to develop tools allowing us to build domain specific data warehouses, which automatically integrate information found on the Web with private information, and information provided by content providers. This work takes place within the RNTL project e.dot that started during 2003. The project is based on XML, and on new services, such as those provided by Xyleme, high level queries and Web monitoring. Experimentation will lead to the construction of a data warehouse on food risk.
This project is a cooperation between the INRA BIA group, the Xyleme start up, and Gemo. BIA was chosen by the Ministry of Agriculture and the Ministry of Research to be the center of computer related skills with regards to the national research program on food risk.
Some recent software developed in Gemo:
ActiveXML: a language and system based on XML documents containing Web service calls. ActiveXML is now in Open Source within the ObjectWeb Forge)
SomeWhere: a P2P infrastructure for semantic mediation.
KadoP: a peer-to-peer platform for warehousing of Web resources .
TreeFinder, Dryade: prototype systems that discover frequent tree patterns within a collection of XML data.
OntoMedia: a prototype for the automatic construction of ontology components, using DTDs, developed within the PICSEL2 project.
XQueC: a prototype for storing and querying compressed XML data.
WebQueL: a multi-criteria filtering tool for Web documents, developed in the setting of the e.dot project.
Acware: a prototype of Web warehouse definition and construction, based on a declarative language, and implemented using ActiveXML
STYX: definition and construction of a generic platform to integrate and query relevant XML resources concerning a Web community.
XSum: an free open-source XML visualization and analysis tool .
One of the reasons for the success of the relational data model was probably its clean theoretical foundations. On the mathematical side it simply consists of relations equipped with first-order logic as a query mechanism. This is accompanied by the equivalent relational algebra which allows evaluation of queries and facilitates optimization issues. Last but not least, there is SQL as an easy-to-use query language which, thanks to its simplicity, evolved de facto as a standard.
Obtaining such a clean foundation for the semistructured data model and XML is still an open issue. We studied XML as well as Active XML, a novel extended variant of XML where part of the documents data is generated dynamically via embedded calls to Web services. We detail further our results, first for XML and then for ActiveXML.
XML Most of the current proposals are based on the tree structure of XML data and make use of the fundamental connection between Monadic-Second-order (MSO) logic and automata on trees. Most of our theoretical work follows this approach.
In we study the precise complexity of testing whether an unranked tree is accepted or not by a tree automata. We deduce from it the precise complexity of checking whether an XML document conforms to a DTD or an XML-schema. We also study, in the same paper, the precise complexity of evaluating queries of XPath and various fragments of XPath (a W3C standard which is in the core of many XML query languages).
ActiveXML (AXML) extends XML by allowing documents where some of the data is given explicitly while other parts are defined only intensionally by means of embedded calls to Web services (see Section ). Each such call is typed using XML schemas. A formal foundation for this new generation of AXML documents and services is presented in . This paper also study the fundamental issues they raise. The focus is on Web services that are (1) monotone and (2) defined declaratively as conjunctive queries over AXML documents. In this case it studies the semantics of documents and queries, the confluence of computations, termination and lazy query evaluation.
In the AXML setting, when a user requests some of the data, the system has to decide whether to materialize some of the intensional data or not and the order of the materialization may affect the final result. In and the general problem is shown to be non computable and an overview of many decidable cases is presented.
In the continuation of our work on discovering frequent trees in huge collections of tree data, we have proposed a novel algorithm (DRYADE) for discovering frequent trees, even in the presence of variations in the nesting of the labels of XML documents. The solution experimented in the Dryade algorithm is to compute closed frequent trees, which has the advantage of providing a compact representation of frequent trees without loss of information. The performance of Dryade has been experimentally tested on artificial medium-size data and also on real data in the setting of the Xyleme project1.
We have extended our work implemented in the ZooM system for clustering semistructured data based on Galois lattices by the definition and the study of new Galois lattices, which we have called alpha Galois lattices, allowing the control of the number of nodes of the lattice and thus the adaptability to users and data. The theoretical foundations of alpha Galois lattices have been presented in .
In the continuation of the work on the automatic construction of a ontology from a set of XML documents corresponding to standardized service specifications in a given domain, we have extended the prototype OntoMedia by functionalities for generating suitable user interfaces from the ontology and for generating wrappers from the XML service specifications .
We have designed the Somewhere PDMS whose data model complies with the W3C recommendations since it is captured by the propositional fragment of the OWL ontology language. Our work on distributed reasoning in propositional logic provides the query-answering algorithm of Somewhere since we have shown that in our data model query answering amounts to a proper prime implicant calculus . This simple data model already scales up to more than a thousand peers in our first benchmarks. Moreover, it guarantees that the query-answering problem is decidable. Somewhere is the basis for the forthcoming MediaD project in collaboration with France Telecom R&D on efficient distributed mediation.
In continuation of our work on query refinement based on the construction of a Galois lattice, we have made several experiments on the OntoRefiner prototype in order to analyze the number and the content of the clusters when there are a lot of answers, and we have implemented several optimizations . Our heuristics exploit the taxonomy of concepts and allow to compute only approximate clusters, which are very close to the real clusters. The use of approximate clusters allow to avoid the closure computation which is the most costly step.
In the context of ActiveXML, we have addressed the problem of ``dangling'' service calls, i.e. service calls that return an error instead of a valid result. We have introduced the notion of service call plan which is composed of a service call and a set of possible alternative service calls fired by specific error events. A prototype has been implemented by Radu Pop during his internship and we are currently studying the semantic description and automatic generation of service call plans as a semantic extension of the ActiveXML model.
Web services can be seen as the building blocks for complex software applications, see, e.g. Microsoft .Net, or BEA Web Logic. We have continued the development of the ActiveXML (AXML, for short) system, a declarative framework that harnesses Web services for data integration, and is put to work in a peer-to-peer architecture - see also http://activexml.net. An overview of the project is now available . The AXML system is centered around XML documents where some of the data is given explicitly while other parts are defined only intensionally by means of embedded calls to Web services. We considered the exchange of such documents between applications, and their distribution and replication among peers.
The issue of query processing in this context is addressed in . Other aspects of ActiveXML have been considered such as the flexibilily of query answering , security and access control
We have performed a detailed performance analysis on the AXML peer implementation for PCs , identifying all elementary cost components associated to a service call activation. Furthermore, we have proposed a new execution model and an associated distributed optimization strategy for the problem of AXML document materialization .
In the field of Web application design, conceptual modeling has already demonstrated its advantages, allowing for declarative specification, easier correctness checks, and automatic deployment, from a high-level model to implemented code. In collaboration with the WebML team from Politecnico di Milano, Italy, we have extended the WebML (www.Webml.org) modeling language with support for Web services and workflow; we demonstrated WebRatio, a commercial tool implementing this approach, in the SIGMOD conference 2004 .
This theme has been driven in Gemo by the RNTL e.dot project. Thematic warehouses integrate data collected from multiple sources and relative to a same application domain described by an ontology. In contrast with mediators, which manage virtual data, warehouses store data extracted from information sources and then integrated with the data already present in the warehouse. In all our work on Thematic Warehouses, a domain ontology is used. It guides data acquisition, extraction and integration. Moreover, as XML is nowadays the new standard for the exchange of data, it has been chosen as the representation language of the warehouse. Our work is motivated by and developed in the context of the e.dot RNTL project which aims at enriching an existing data warehouse (in the domain of food risk assessment) with data from the Web (http://www.lri.fr/~cr/images/SchemaEdot.JPG). In e.dot, the domain ontology describes the contents of the warehouse in two parts: a relational database and data represented in the conceptual graph formalism. The ontology plays the role of the schema of the warehouse. It is a set of concepts organized in a hierarchy with their synonyms in French and their translation in English. The organisation and the materialisation of the warehouse is specified in a high-level model specific to thematic warehouses and with all operations invoked through Web services (ACWARE Active Content Warehouse). These specifications are executed by ActiveXML, a framework developed by Gemo to deal with embedded Web services calls inside Web documents .
Our work combines a search engine such as Google or a web crawler (such as that of Xyleme) with a filtering tool that can distinguish, among the possible thousands of web pages returned by Google or the Xyleme crawler, those that really contain useful data for the warehouse. In the first e.dot experiments, it was shown that guiding the search through the Web by keywords extracted from the domain ontology was not precise enough to guarantee that the returned Web pages were relevant to the topic of the warehouse. Our approach for designing a filtering tool is generic and declarative. We have defined and implemented a query language, called WebQL, which enables the combination of different criteria for specifying the Web pages of interest. Those criteria allow for combining content and structure of searched documents. For example, using WebQL, it is possible for the warehouse administrator to specify that he is interested in HTML pages containing tables with titles including a given keyword. The evaluation of a WebQL query does not provide a ``yes or no'' answer but a matching coefficient, which is the basis for ranking the pages that are filtered. We plan to integrate WebQL within a focused crawling tool.
We want to populate the thematic data warehouse with data found in tables of HTML or PDF documents because tables often contain relevant and synthetic data. We explicitly separate tasks that are specific to the format of a web source, e.g., HTML, from the tasks that are independent of any source but specific to the domain. Thus, we first automatically transform the various tables into a generic XML representation called XTab. Because of the semantic heterogeneity among sources, extracting data from Web pages is insufficient to support integration. Data has to be organized in a different way with a different vocabulary, i.e. we have to find an XML representation where most of the values and tags belong to the ontology. In our approach, we want this transformation to be as automatic and flexible as possible, only driven by the ontology and the way the data has been structured in the original table. Thus, we have defined a Document Type Definition named SML (Semantic Markup Language) that can automatically be generated using the ontology and can deal with additional or incomplete information in a semantic relation, ambiguities or possible interpretation errors. This transformation has been partly implemented and experimented on real data from the e.dot project. The extracted data will be exploited by an interrogation engine named MIEL++, which is developed by the INAPG team.
Our aim is to allow a user to interrogate with a single query several data sources related to the same topic but annotated with terms from multiple ontologies. In order to be able to query the other sources, mappings between terms of the different ontologies are needed. We propose to compute these mappings once and for all and to store them in the warehouse in order to be used by the query engine. The identification of the mappings combines both syntactic and semantic comparison techniques, i.e. name-based matches and structure-level matches. This research work is being applied in the e.dot project with two ontologies: Sym'Previus and Com'Base. A first tool supporting automatic syntactic mapping techniques has been developed.
We are also developing a flexible and generic approach, which would let us specify in a declarative way the information necessary to create and enrich a thematic warehouse. We also want to simplify the acquisition of the documents that should be stored in the warehouse from the Web, monitor this warehouse, and organize the information it contains, for future querying. We have begun a first experimental prototype, based on the ActiveXML language. To this end, we have programmed a library of Web services useful in order to construct a Web warehouse.
The problem of XML query evaluation still poses significant challenges. In particular, the complexity of the XQuery language, standardized by the W3C, makes it very difficult to devise efficient storage and optimization strategies. We have finalized our work on the XQueC compressed XML database system , and we have re-designed and isolated a set of XML database functionalities into our GeX prototype . GeX uses an efficient path-driven partitioning strategy, allowing it to process complex structural XML queries. Going beyond the framework of a single storage model, we have started to work on a generic XQuery optimization model, capable of coping with varied and changing persistent storage modules. This capacity is crucial in order to support efficient data access, e.g., indexes and materialized views. The MS internship and beginning of the PhD thesis of A. Arion focused on this topic ; a generic optimizer prototype is currently under development . As part of our work in this area, we have organized, in cooperation with Y.Papakonstantinou (UCSD, USA) a workshop centered on XQuery, in cooperation with ACM SIGMOD .
We have designed and implemented KadoP, a peer-to-peer platform for building and managing warehouses of Web resources . KadoP relies on a Distributed Hash Table implementation (namely, FreePastry) to keep the network of peers connected, and to build a shared global resource index, and on the ActiveXML platform to store, query, and maintain the index. Furthermore, KadoP is able to process simple queries carrying over resources distributed in the whole network.
A main goal is to be able to index not only extensional XML data but also intensional one and in particular Web services.
As an application of the work around KadoP, we have applied our thematic Web warehousing approach to the study of the XQuery standardization process, within the framework of the ACI ``Normes et Politiques Publiques''. We have devised a conceptual model for the actor interactions taking place on the public XQuery mailing list, loaded this list into our warehouse, and performed some preliminary analysis .
As another application, we are participating in the INEX (Initiative for the Evaluation of XML Information Retrieval), more precisely in its Heterogeneous track. This track focuses on IR-style querying over a corpus of heterogeneous XML sources from the domain of CS bibliographic data. We have applied our thematic Web warehousing approach to construct an unified conceptual schema, and a set of mappings from individual schemas to the unified one, which solve the schema heterogeneity problem .
Gemo has had technical meetings in 2004 with several industrial partners, in particular France Telecom R&D, Xyleme, Mandrakesoft, eNetshare, and Exalead, as well as national organizations, in particular, Institut National de Recherche en Agronomie and Bibliothèque Nationale de France.
This project is the continuation of PICSEL1, which focused on designing a declarative environment for building ontology-based mediators. It aims at scaling up to the Web the mediator approach that has been implemented in PICSEL1. The goal is to facilitate the automatic construction of a mediated schema over several XML sources described by DTDs and related to a same domain. A prototype (OntoMedia) has been developed, which extracts ontology components automatically from a set of DTDs. In PICSEL2, we also develop methods initiated in PICSEL1 for cooperative query answering.
The goal of the EDOS Project that started in 2004 is to boost quality and productivity in software development in the field of Open Source software. Typically, most recent Linux operating systems comprise thousands of individual packages. This makes putting together such a system a task of daunting difficulty; and the short release cycles traditionally practiced in Open Source software mean that this task has to be constantly repeated. And this is the case not just for Linux, but for any large, modular software project.
In this context, the Gemo group will focus on the management of distributed information (the software releases).
Participants include academic labs INRIA (Gemo and Cristal), the U. of Paris 7, Tel-Aviv, Geneva, Zürich and Torino; and private companies Edge-IT (a Mandrakesoft subsidiary), Nuxeo, Nexedi (France) and SOT (Finland).
The goal of the e.dot project is to develop an XML warehouse for information concerning food safety risk analysis and management. It is composed of Gemo, the BIA Group of the Institut National de Recherche en Agronomie (INRA) and the Xyleme company.
One key point of this project is that for constituting the XML warehouse, we are guided by a pre-existing ontology that was designed by INRA people as the uniform schema of their data, in order to acquire relevant documents from the Web, transform them into semantically enriched XML documents (using ontology terms as element tags). We started by specifying and implementing the different functionalities and modules that are necessary to meet the application needs and to take advantage of existing application knowledge (ontologies) and data. These modules have been packaged and integrated in an open, service-oriented architecture composed of e.dot services for data acquisition and semantic enrichment, an e.dot XML warehouse for data storage (ActiveXML and Xyleme), and a graphical query interface (MIEL++) integrating this warehouse with several other pre-existing data sources. The service integration and warehouse definition process has been guided by using the SPIN approach and we are currently studying a new and more performant solution for the storage and querying of ActiveXML documents based on an integration of the ActiveXML system and the Xyleme product.
In France, close links exist with groups at Orsay (databases, N. Bidoit; bio-informatics, C. Froidevaux, C. Rouveirol; machine learning, M. Sebag), with the Cedric Group at CNAM-Paris; some INRIA groups (Atlas, P. Valduriez, DistribCom, A. Benveniste, at INRIA-Bretagne); the BIA group at INRA (O. Haemmerle, P. Buche, C. Dervin), the LISI of the University of Lyon 1 (M. Hacid), and the LIRMM of the University of Montpellier (M. Chein, M-L. Mugnier).
This project is funded by the ACI (Action Concertée Incitative) Masses de Données. It is a joint project with Patrick Gallinari's group of LIP6 (University of Paris 6) and Remi Gilleron's Mostrare group.
The goal of this project is to study fundamental problems raised by modern information retrieval and to determine novel solutions to solve these problems. In particular, we want to build tools for retrieving and extracting information, which fully and jointly exploit the structure and contents of the XML documents. The distinguishing feature of our approach is to use machine-learning techniques for building flexible and robust tools applicable to large corpora of structured documents, which are possibly heterogeneous, varied and dynamic.
This project on Massive Data Management in Peer-to-Peer Systems is funded by the ACI Masses de Données. MDP2P is a joint project with the Atlas, Paris and TexMex teams from INRIA-Bretagne. The goal of this project is to provide efficient data management tools in a peer-to-peer architecture. Our work in this project focused on harnessing the expressive power of ActiveXML to devise robust, large-scale platforms for sharing XML resources. Several of our results concerning distributed data management with ActiveXML are part of our work in the MDP2P project . We have performed an extensive performance study of the ActiveXML basic platform for PCs , and have proposed novel execution methods and optimization algorithms for computations over many peers . In parallel, we have designed and built KadoP, a platform for large-scale sharing of Web resources such as semi-structured documents and Web services . KadoP combines the power of ActiveXML and that of a Distributed Hash Table implementation (FreePastry), to perform efficient resource lookup and querying.
TraLaLa stands for XML Transformation Languages: logic and applications. It is funded by the ACI Masses de Données and has just started. The setting is the integration and manipulation of massive data in XML format. We are interested more specifically in the programming and querying languages aspects: expressivity, typing, optimization. We are also interested in studying how this can be done in a context where documents are compressed or in a streaming scenario. The project is funded for three years. Its home page can be found at: http://www.cduce.org/tralala.html.
This project has just started, as a collaboration with Benjamin Nguyen (University of Versailles) and with several political scientists (F.-X. Dudouet from University of Paris X). Our purpose is to investigate the process of drawing up Information Technology standards and regulations, to understand how the process proceeds, who are the actors involved, and which are the actual mechanisms for setting up a standard. Our task in this project is to design and prototype an architecture for the construction of a semistructured data warehouse, used as a support for verifying the social scientists' hypothesis. Thus, our participation in this contract draws on our experience with the SPIN and KadoP prototype. We have taken the W3C XQuery standardization working group as an example, and have performed a preliminary study of the interactions on the public mailing list of the standard committee .
In Europe, close links exist with Tel Aviv University (T. Milo), University of Marburg (T. Schwentick), University of Athens (M. Vazirgiannis), University of Madrid (A. Gomez-Perez), University of Manchester (I. Horrocks), University of Rome (M. Lenzerini) and Politecnico di Milano (S. Ceri).
Particular projects that we conduct are detailed next.
This year is the second year of a PAI-Procope project with the database group of Bernhard Seeger and Thomas Schwentick at Marburg University, Germany. The project is expected to last until the end of 2005. Its goal is to generate interactions between theory and practice in the context of systems for semi-structured data. More specifically we would like to find out whether we could develop automata- and logic-based methods for XML query evaluation and optimization.
Close links exist with the Hebrew University (C. Beeri) and the University of Tel-Aviv (T. Milo returned to Israel in 2004 after a long visit in the group).
In the US, close links also exist with the Stanford University (J. Widom), AT&T (S.Amer-Yahia), University of Washington (A. Halevy), University of Rutgers (A. Borgida), University of Toronto (A. Mendelzon and L. Libkin).
Since 2003, Gemo and the data management group at the University of California at San Diego (V. Vianu, A. Deutch, Y. Papakonstantinou) form an associated team funded by INRIA International. This association is expected to last at least three years. The two groups met in Paris in June for a three days workshop, XIME-P, co-organized by Ioana Manolescu and Yannis Papakonstantinou (UCSD). Moreover three seniors and one Ph.d students from UCSD came to visit Gemo in Paris and stayed from one week to three months. The home page of GemSaD can be found at http://www-rocq.inria.fr/~segoufin/GEMSAD/
GemSad is also supported by the National Science Foundation until 2006.
This year the following professors visited Verso:
Michael Benedikt, Lucent Research Lab (June)
Tova Milo, professor at the University of Tel-Aviv (till June)
Victor Vianu, professor, UC San Diego (June to August)
The following students came for internships in the group: Zoe Abrams [Stanford U.; 3 months]; Alan Nash [UCSD; 1 month]; Antonella Poggi [U. Roma; 2 months; joint European PhD]; Gabriella Ruberg [Federal University of Rio; 6 months]; Cristina Sirangelo [U. of Calabria; 8 months].
Nicolaas Ruberg also spent an internship in the group (6 months in 2004), sent by the Bank NDES of Brazil.
Serge Abiteboul, Sophie Cluet, Michel Scholl and Vassilis Christophides are the 2004 recipients of SIGMOD 2004 Test of Time Award for their paper "From Structured Documents to Novel Query Facilities" in SIGMOD 94. Sophie (ex manager of Verso) is now Director of INRIA Rocquencourt; Michel (ex manager of Verso) is now Professor at CNAM-Paris; Vassilis (ex PhD student of Verso) in now Professor at the U. of Crete.
Marie-Christine Rousset has been General Co-chair of FQAS 2004 (Flexible Query Answering Systems) and Tutorial chair of ECAI 2004 (European Conference on Artificial Intelligence).
Ioana Manolescu has organized and been Co-Program chair of XIME-P, the First International Workshop on XQuery Implementation, Experience and Perspectives in cooperation with ACM SIGMOD, 2004.
Members of the project have participated in program committees:
S. Abiteboul
Workshop on XML High performance, W3C, 2004
Web and database workshop, 2004
International Workshop on Web query language, 2004
ACM Symposium on Applied Computing (ACM SAC), Cyprus, 2004
World Wide Web Conference 2004, WWW2004
World Wide Web Conference 2005, WWW2005
IEEE International Conference on Web Services (ICWS '05), Orlando 2005
I. Manolescu
VLDB (International Conference on Very Large Databases) 2004
ICDE (International Conference on Data Engineering) 2004
International SIGMOD conference 2004
International AIMSA (Artifficial Intelligence: Methods, Systems and Architectures) Conference 2004
International ICDE/EDBT PhD workshop 2004
Journees de Bases de Donnees Avancees" (BDA) 2004, the French database conference
"Simposio Brasileiro de Banco de Dados" (SBBD) 2004, the Brasilian database conference
ACM Workshop on Web Information and Data Management (WIDM) 2004
C. Reynaud
International Conference on Knowledge Engineering and Management (EKAW), 2004.
3rd International Semantic Web Conference (ISWC), 2004.
Workshop on Knowledge Management and the Semantic Web, International Conference on Knowledge Engineering and Management , 2004.
Journée Web Sémantique Médical, Rouen, 2004.
Atelier Modélisation des connaissances, EGC, Janvier 2005.
Ingénierie des Connaissances, Mai-Juin 2005.
M-C. Rousset
KR 2004 (International Conference on Principles of Knowledge Representation and Reasoning).
SIGMOD 2004 (ACM International Conference on Management of Data).
Congrés Francophone de Reconnaissances des Formes et Intelligence Artificielle (RFIA), 2004
L. Segoufin
ACM International Symposium on Principles of Database Systems (PODS), Paris, France, June 2004
Serge Abiteboul has been invited to present Distributed information management with XML and Web services at ETAPS04, the European Joint Conferences on Theory and Practice of Software, in Barcelona. he has also been invited to present ActiveXML and Active Query Answers at FQAS 2004, International Conference on Flexible Query Answer, in Lyon. He has also been invited to present ActiveXML, Security and Access Control at SBBD 2004, XIX Brazilian Symposium on Databases, in Brasilia.
Ioana Manolescu presented an invited tutorial on XML Query Processing: Storage and Query Model Interplay at the International EDBT summer school in databases, and at the SBBD conference, 2004. She also presented an invited tutorial on XML query processing and distributed applications at the DRUIDE 2004 summer school, Domaine de Port-aux-Rocs, France, May 2004. She presented also this second tutorial at the AlgoDis summer school in Porquerolles, France, September 2004.
Chantal Reynaud has been invited to present Semantic integration of Heterogeneous Data at Topical session Workshop, IFIP World Congress (Toulouse, Août 2004).
Marie-Christine Rousset presented an invited tutorial on Médiation sémantique dans un réseau Pair-à-Pair at Summer School DistRibUtIon de Données à grande Echelle (DRUIDE), Le Croisic, France, May 2004, She also had an invited presentation on Small Can Be Beautiful in the Semantic Web at the International Web Semantic Conference (ISWC 2004), Hiroshima, Jaan, November 2004.
B. Amann
Revue d'Intelligence Artificielle (RIA)
Revue Ingénierie des Systèmes d'Information, Numéro Spécial "Les services web, théories et applications" (2005)
C. Reynaud
JEDAI (Journal Electronique d'IA de l'AFIA)
Revue Information - Interaction - Intelligence (I3 )
M-C. Rousset
ACM Transactions on Internet Technology (TOIT)
AI Communications (AICOM)
Electronic Transactions on Artificial Intelligence ( ETAI) (for the areas: Concept-based Knowledge Representation and Semantic Web).
Revue Information - Interaction - Intelligence (I3 )