The CNR-IEI team will contribute to the development of a logic for MIR based on terminological logic. In doing this, the team will bring to the FERMI consortium its experience on conceptual multimedia document modelling based on terminological logic. The CNR-IEI team will also work on the model for multimedia data, on the basis of its experience on image modelling and retrieval.
The CNR-IEI team is active since 1985 in the field of the conceptual modelling of multimedia documents, image retrieval, and, more recently, logic-based information retrieval.
The CNR-IEI team will bring in the project the experience gained by its participation to the ESPRIT project MULTOS (project no 28, 1985-1990). In this project, research and prototyping activities were carried out on certain innovative aspects of multimedia document systems:
From 1992, CNR-IEI is coordinating of the ESPRIT BRA MIRO Working Group (no. 6576, 1992-1995). The Group is searching new methods and techniques for information storage and retrieval so that all types of media can be handled in an integrated manner through adaptive interaction with the user. A major aim is to develop example logics in the context of object-oriented approaches to data models, and storage structures within which new retrieval models, based on formal logic and a theory of uncertainty, will address the claim that retrieval can be viewed as a form of inference. The CNR-IEI team contribution is the development of a logic that extends a suitable terminological logic with probabilistic inference and content representation of text and images.
The CNR-IEI team has also been involved in the ESPRIT BRA project Formally Integrated Data Environment (FIDE1- project no. 3070, 1989-1991). The goal of the FIDE1 action was to develop the foundations for an integrated data environment through advances in type systems, programming languages, database systems and conceptual modelling. The individual technologies have been studied to analyse the precise functions they provide. These showed both overlap and inconsistency. The action sought to eliminate the inconsistency and hence make application system building and maintenance much more economic. At the same time FIDE1 developed a support architecture which aimed at being significantly more efficient having removed unnecessary and often conflicting overlap. The CNR-IEI team contribution regarded the design of a high performance object store and the application of formal specification techniques to the design of database applications.
The team is now involved in the continuation of the FIDE1 action, that is the FIDE2 Esprit Basic Research Action (no. 6309, 1992-1995). FIDE2 will extend the work of the first FIDE1 project into fuller integration. Such integration will provide significant economy in the construction of sophisticated long-lived applications. This will be achieved by imposing consistency on the supporting technologies. Specifically, a consistent framework will house a consistent set of primitives supporting precise data description, persistent data management, concurrency, distribution and recovery. This will be supported via persistent object stores and used via persistent and database programming languages.
The team is a member of the ESPRIT Network of Excellence no. 6606, ``Information and Data on Open Media for Networks of Users - IDOMENEUS''.
Costantino Thanos
Costantino Thanos (born 1942) holds a Doctorate in Electrical Engineering from the University of Pisa.
Since 1970 he has worked at CNR-IEI and is currently head of the ``Advanced Information Processing Techniques'' department. His research activities include: Project Coordinator of an Italian nation wide project on Distributed Data Bases: DATANET (1980-1984); Coordinator of the Working Group ``Multimedia Information Retrieval- MIRO'' (ESPRIT Basic Research Actions Working Group No.6576, 1992-1994); Responsible for CNR teams participating in a number of international projects: (a) Evaluation and Implementation of Database Systems, (b) Performance Evaluation of Concurrency Control Mechanisms in a System for Distributed Databases using Broadcast Networks employing Satellites, (c) Architecture for Heterogeneous European Distributed Data Bases, (d) Mixed Mode Message Filing System (1984 - ESPRIT pilot project), (e) Multimedia Office Server (ESPRIT project No. 28, 1985-1989), (f) Tools for Designing Office Information Systems (ESPRIT project No. 813, 1986-1988), (g) Construction and Management of Distributed Office Systems (ESPRIT project No. 834, 1986-1988), (h) Formally Integrated Data Environment (ESPRIT Basic Research Actions project No. 3070, 1989-1992).
He has also been appointed: Associate Editor of the international journal ``The Computer Journal'' published by the Oxford University Press; Member of the Editorial Board of the international journal ``Computer Standards &Interfaces'' published by the North Holland Publishing Company; Member of the Executive Committee of the ``European Research consortium for Informatics and Mathematics - ERCIM''; Member of the evaluation team for the ESPRIT Program- pilot phase, phase II, phase III in the Office and Business Systems Area; Member of the evaluation team for the Specific Programme and Technology Development in the field of Telematic Systems of General Interest, area 6 - Linguistic Research and Engineering (LRE) (1992); Member of the Basic Research Working Group-Computer Science (ESPRIT Program) (1990,1992); Reviewer of ESPRIT projects: ADKMS,.INDOC, INBAS, TOOTSI, MIPS, DOCS; Member of the PC of the International Conference ``Very Large Data Bases'' in 1983, 1984, 1985, 1986, 1987, 1989, 1990; Member of the PC of the International Conference ``Conference on Advanced Information Systems Engineering- CAiSE'' in 1990, 1991, 1992, 1993; Member of the PC of the International Conference ''International Conference on Multimedia Information Systems'', Singapore, 1991; Member of the PC of the International Conference ``Visual Database Systems'', Budapest, 1991; Member of the PC of the International Conference ``Database and Expert Systems Applications-DEXA'' 1991, 1992, 1993.
Recent Publications:
Foundations of knowledge Base Management. Springer-Verlag Topics in Information Systems, Heidelberg, FRG, 1989 (edited with J.W. Schmidt).
Multimedia Office Filing: The MULTOS Approach. North Holland Series in Human Factors in Information Technology No. 6, Amsterdam, NL, 1990 (editor).
Conceptual Document Modelling and Retrieval. Computers Standards and Interfaces, 11(3):195-213, 1990/1991 (with C. Meghini and F. Rabitti).
Conceptual Modeling of Multimedia Documents. IEEE Computer, 24(10):23-30, 1991 (with C. Meghini and F. Rabitti).
A Model of Information Retrieval based on a Terminological Logic. To appear in Proceedings of SIGIR-93, 16th International Conference on Research and Development in Information Retrieval, Pittsburg, PA, 1993 (with C. Meghini, F. Sebastiani and U. Straccia).
Carlo Meghini
Carlo Meghini (born 1956) graduated in Computer Science at the University of Pisa in 1979; until the end of 1979 he worked at the Computer Science Department of the University of Pisa, and in 1981 at the Mathematics Department, supported by an IBM Research Grant.
Since 1984 he has been a member of the research staff at CNR-IEI in Pisa, working in the area of distributed database management systems, conceptual modelling and logic-based retrieval. From September 1984 to December 1985 he was a visiting Research Scientist at the Computer Science Department of University of where he worked in the Knowledge Representation Group led by Prof. John Mylopoulos on topics concerning the use of Knowledge Representation in the design of Information Systems and Conceptual Modelling.
Since 1986 he has been involved in several ESPRIT projects, funded by the European Economic Community, in the area of office systems (MULTOS, TODOS, COMANDOS). He has also participated in the FIDE project, as part of the ESPRIT II Basic Research program and is currently responsible of the modelling retrieval team in the MIRO Working Group (No. 6576).
Recent Publications:
The Complexity of Operations on a Fragmented Relation. ACM Transactions on Database Systems, 16(1):56-87, 1991 (with C. Thanos).
Conceptual Document Modelling and Retrieval. Computers Standards and Interfaces, 11(3):195-213, 1990/1991 (with C. Thanos and F. Rabitti).
Conceptual Modeling of Multimedia Documents. IEEE Computer, 24(10):23-30, 1991 (with C. Thanos and F. Rabitti).
Multimedia Document Handling. In Proceedings of the International Conference on Multimedia Computer and Communication: Technology, Application and Enterprise, (invited paper), Bombay-India, 1992 (with C. Thanos).
A Model of Information Retrieval based on a Terminological Logic. To appear in Proceedings of SIGIR-93, 16th International Conference on Research and Development in Information Retrieval, Pittsburg, PA, 1993 (with C. Thanos, F. Sebastiani and U. Straccia).
The GU-CS team will bring to the project its considerable expertise in developing retrieval models based on logics and probability theory. It will lead the probabilistic research line and contribute significantly to the experimentation. There is a long history of successful IR experimentation associated with the researchers at Glasgow.
The GU-CS esearch group has a long history in Information Retrieval research. For example, the first complete version of the Probablistic Model was formulated in the seventies by members of the Glasgow team then at Cambridge University. Furthermore, over the years a significant experimental methodology has been developed by the researchers now in Glasgow. Much of this early work came to fruition in MINSTREL (New information models for office filing and retrieval, Esprit project 59) in which IR techniques were integrated with data base models without the help of any underlying framework. This lack was recognised as a limitation. In the last few years IR research at Glasgow has concentrated on constructing a formal framework for specifying new IR models. There are a number of collaborative projects in the department concerned with the storage, manipulation, display and retrieval of multimedia objects. One of these is KWICK (project 2466) and another is SHAPE (project 5398).
There are a number of researchers working in the area of multi-media and logics for IR. One is researching into providing access, through free text queries, to mixed media information held in a hypermedia document base. Another is investigating the use of simple Natural Language Processing techniques to improve the effectiveness of IR systems, specifically using online dictionaries to work out the sense of words in texts. A third researcher is studying the use of advanced user interfaces to information retrieval servers available over wide area networks. A fourth is concerned with the connection between situation semantics and conditional information. A fifth has been investigating the use of networks to represent the propagation of uncertainty in the computation of the probability of relevance for documents.
GU-CS has been involved in the ESPRIT funded projects Comandos (project 834), and FIDE (project 3070), and is currently involved in the Comandos 2 (project 2071), IMIS (project 6548), and Semantique (3124) ESPRIT funded projects, and in the MIRO (project 6576) and SemantiqueII (project 6809) working groups. It has internationally funded collaboration with sites in France, Germany, and Italy through the British Council and also have international funding through the Media Trust. It is currently involved in three Science and Engineering Research Council funded projects: Configuring Data, TAU, and AQUA/DSA. It also has Royal Society and Digital Corporation funding for research work.
The department has an active involvement in the IDOMENEUS network of excellence and is the prime contractor for the FIDE2 project. This together with work on visualisation of data, object oriented approaches to IR, and computer animation provide an excellent wider context for the FERMI work at Glasgow.
Cornelis Joost van Rijsbergen
Cornelis Joost van Rijsbergen (born l943) holds a BSc and a Dip NAAC from the University of Western Australia, and a PhD from the University of Cambridge.
Appointments: l966-68, Tutor in Mathematics, Mathematics Department University of Western Australia; l969-72, Senior Research Officer King's College Research Centre, Cambridge; l973-75 Lecturer Department of Computer Science Monash University, Melbourne, Australia; l975-79 Royal Society Scientific Computer Laboratory Information Research Fellow, University of Cambridge; l978 Spring Visiting Professor, School of Library and Information Studies, University of California at Berkeley; l980-86 Professor and Head of Department of Computer Science, University College Dublin, Ireland; l986-89 Professor, Department of Computing Science, University of Glasgow, Scotland; 1990-93 Head of Department, Department of Computing Science, University of Glasgow, Scotland.
Professional Activities: Associate Editor (Europe) for Information Processing and Management (including Information Technology), Pergamon Press; Editor-in-Chief for The Computer Journal Oxford University Press; Editor of Intelligent Systems Engineering (IEE); General Series Editor for Cambridge Tracts in Theoretical Computer Science, Cambridge University Press; General Series Editor for Workshops in Computing, Springer-Verlag; General Series Editor for Distinguished Dissertations in Computer Science, Cambridge University Press; Advisor to GMD; Director of Itext Ltd, Inforythmics Ltd (IT mutimedia companies).
Recent Publications:
NRT (News Retrieval Tool). Electronic Publishing: Origination, Dissemination and Design, 4(4):205-217, 1991 (with M. Sanderson).
Hypermedia and Free Text Retrieval. Information Processing and Management, 11, 1992 (with M.D. Dunlop).
Probabilistic Retrieval Revisited. The Computer Journal, 35(3):291-298, 1992.
A Logical Model of Information Retrieval based on Situation Theory. In Proceedings of BCS 14th Information Retrieval Colloquium, Lancaster, UK, 1992 (with M. Lalmas).
The state of information retrieval: logic and information. The Computer Bulletin, 5(1):18-20, 1993.
Mark David Dunlop
Mark David Dunlop holds a Bachelor of Science, honours of the first class, Glasgow 1988, and is a Doctor of Philosophy in Computing Science, Glasgow 1991.
In 1988-1991 he has been a PhD Student in Glasgow supervised by Prof. C.J. van Rijsbergen, and in 1991-1993 a Lecturer in Computing Science at the University of Glasgow. He is currently an ``academic'' in Computing Science at Glasgow.
Recent Publications:
Multimedia Information Retrieval, Ph.D. Thesis available as Research Report 91/R21, Computing Science Department, University of Glasgow, UK, 1991.
Hypermedia and free text retrieval. In Proceedings of IEE Computing and Control Division Colloquium on Hypertext, Digest no. 1990/142, IEE Savoy Place, London, UK, 1990.
Access methods to non-textual documents. In Proceedings of the 14th International Conference on Research and Developments in Information Retrieval, Singapore, 1991 (with C.J. van Rijsbergen).
Hypermedia and probabilistic retrieval". In Proceedings of RAIO 91, Conference on Intelligent Text and Image Processing, Catalonia, Spain, 1991 (with C.J. van Rijsbergen). A highly revised and updated version of this paper is to be published in Information Processing and Management (should be available within the first half of 1993).
Hypermedia and free text retrieval. In K.W. Waite, editor, Current Human-Computer Interaction: An Anthology of Recent Papers (Volume II), pages 67-88, Research Report GIST-91-1, Computing Science, University of Glasgow, UK, 1991 (with C.J. van Rijsbergen).
Due to the background described below, our role in this project would be mainly focused on the problem of modeling multimedia data (either structured or unstructured), and on the development of a logic-based retrieval model for this data. The first topic concerns the design of appropriate indexing languages for multimedia data that encompass both semantic and structural aspects of this data, and we would take the responsability of the corresponding work task in the project. We would address this topic through our previous experience on modeling the semantic content of textual data (based on conceptual graphs), extend it to image and graphics media, and to the modeling of structural properties of multimedia data. We would also bring in our experience and, in collaboration with the other partners, participate in the design of a new logic-based retrieval model and contributein its experimentation in using our existing platform and multimedia data (see below).
Since 1983, most of the activities of LGI-IMAG are related to the design of precision-oriented information retrieval systems dedicated to domain-restricted applications such as technical documentation (IOTA project [CDKB86]), medicine (RIME project [Ber90]) and software engineering (ELEN project [CC92]). These systems are based on refined retrieval models based on high-level representation of the semantic content of documents and queries, and involve inference mechanisms as support for demonstrating the relevance of retrieved documents.Through these developments its contribution to the domain have been both theoretical and experimental. Its contribution to basic research has been first aimed to the integration of new approaches such as Artificial Intelligence and Natural Language Processing which were most promising to improve the retrieval process of textual documents. These researches have further induced development of new Retrieval Models that can embody most of these various aspects in a unique, formally assessed framework. A significant result has been the design of a retrieval model based on fuzzy modal logic which has been experimented on multimedia and on software engineering applications (see below RIME and ELEN projects). This model is a first attempt to implement the underlying principle of uncertain implication between documents and queries proposed by van Rijsbergen [vR86b]. All these projects have produced prototype systems based on advanced retrieval models and on high-level representation of the semantic content of documents and queries in order to fulfill the precision paradigm.
As a conclusion LGI-IMAG has developed a considerable experience in the design and in the experimentation of retrieval models able to cope with complex, multimedia data. We think that this experience, mainly centered on logic-based approaches to multimedia information retrieval, will be of central importance in the achievement of the goals assigned to the FERMI proposal.
Marie-France Bruandet
Marie-France Bruandet (born 1944) holds a Licence es Mathematiques (1967), a Diplome d'Etudes Approfondies en Informatique (Grenoble, 1968), a Doctorat de 3eme cycle en Informatique de l'Universite de Grenoble, and a Diplome d'Habilitation a Diriger des Recherches de l' Université Joseph Fourier de Grenoble (1990).
Since 1992 she is a Professor in Computer Science at the Université Joseph Fourier of Grenoble. Since 1978 she is Associated professor in Computer Science at the Université des Sciences Sociales in Grenoble.
Major University and Department activities: chair of the research project Aristote in the Laboratoire de Genie Informatique; director of MIAG (Maitrise d'Informatique Appliquee a la Gestion)
Major Professional Service: referee of the Information Processing and Management Journal; member of the program comittee of the 11th ACM-SIGIR International Conference on Research and Developments in Information Retrieval (Grenoble, June 13-15, 1988) and Publicity Chairwoman; member of the Organizing Committee of the RIAO85 international Conference (Recherche d'Information Assitee par Ordinateur), Grenoble, March 1985.
Recent Publications:
Outline of a knowledge base model for an Intelligent Information Retrieval System. it Information Processing and Management, 25(1):89-115, 1989.
Domain Knowledge Acquisition for an Intelligent Information Retrieval System: Stratregies and Tools. In G. Gouarderes, J. Liebowitz, M. White (eds.), Proceedings of EXPERSYS-90, Expert systems applications pages 231-235, IITT-International, 1990.
A Hypertext Database Model for Information Management in Software Engineering. In Proceedings of DEXA 90, International Conference on Database and Expert Systems Applications, Wien, Austria, pages 69-75, 1990 (with S. Jarwa).
Elen prototype: an Active Hypertext System for Document Management in Software Engineering. In Proceedings of DEXA 92, International Conference on Database and Expert Systems Applications, Valencia, Spain, 1992 (with S. Jarwa).
Matching Objects: One Step. In Proceedings of ICCI, International Conference on Computing and Information. Sudbury, Ontario, 1993 (with P. Mulhem).
Yves Chiaramella
Yves Chiaramella (born 1945) is an Ingenieur de l'Ecole Nationale Superieure des Arts et Metiers (1969), and holds a Diplome d'Etudes Approfondies en Informatique (Grenoble, 1970) and a Doctorat d'Etat en Informatique, Universite de Grenoble (1981)
Since 1983 he is a Professor in Computer Science at the Université Joseph Fourier de Grenoble, and since 1976 he is an Assistant professor in Computer Science at the Université des Sciences Sociales in Grenoble.
Major University and Department activities: director of the Laboratoire de Genie Informatique, Grenoble (since 1989); member of the University Board since 1986.
Major Professional Service: member of the Computer Journal Editorial Board; member of the Information Processing and Management Editorial Board; member of the Intelligent Systems Engineering Editorial Board; chairman of the 11th ACM-SIGIR International Conference on Research and Developments in Information Retrieval (Grenoble, June 13-15, 1988); chairman of the Organizing Committee of the RIAO85 international Conference (Recherche d'Information Assitee par Ordinateur), Grenoble, March 1985; member of the ACM-SIGIR International Conference's Program Committee since 1986.
Recent Publications:
About Retrieval Models and Logic. The Computer Journal. Special issue on Information Retrieval, 35(3): 233-242, 1992, (with J.P. Chevallet).
A retrieval model based on an extended modal logic and its application to the RIME experimental approach. In Proceedings of the 13th International Conference on Research and Developments in Information Retrieval, pages 25-43, Brussels, Belgium, 1990 (with J. Nie).
Indexing medical reports in a multimedia environment: the RIME experimental approach. In Proceedings of SIGIR-89, 21th ACM Conference on Research and Development in Information Retrieval, pages 25-28, Boston, MA, 1989, (with C. Berrut).
A prototype of an Intelligent System for Information Retrieval: IOTA. Information Processing and Management, 23(4):285-303, 1987, (with B. Defude).
IOTA: A prototype of an information retrieval system. In Proceedings of SIGIR-86, 9th ACM Conference on research and Development in Information Retrieval, pages 207-213, Pisa, I, 1986 (with B. Defude, D. Kerkouba, and M.-F. Bruandet).
Based on its strong background in probabilistic retrieval theory and its experience in IR experimentation, the UNIDO-CS group will mainly contribute to the development of a theory of uncertainty and to the experimentation work part.
The UNIDO-CS research group is active in the field of probabilistic retrieval models and the integration of database and IR systems.
The leader of the group has worked in several projects concerned with the development of the automatic system AIR/X which performs indexing with a set of descriptors from a controlled vocabulary. For this purpose, a large indexing dictionary containing pairs of text terms and related descriptors was constructed automatically by analysing manually indexed documents. When indexing new documents, descriptors related to the terms in the text are looked up in the dictionary. After collecting all information leading to the same descriptor, a probabilistic classification procedure assigns an indexing weight. The AIR/X system is applied successfully in the input routine of a large physics database.
For retrieval, different models for using the probabilistic indexing weights were developed, depending on the fact whether query-specific relevance feedback data is available or not. As an alternative to this approach, complex query-document relationships were investigated. Here the indexing dictionary supplies the information for finding possibly relevant documents for a query consisting of descriptors. So the uncertain inference process leads from the document terms to the descriptors and further to the query.
The major contribution of this work to the field of probabilistic IR was the introduction of machine learning concepts, namely abstraction mechanisms and learning algorithms. Whereas former approaches could exploit relevance feedback data only for the same query, the new concepts allow for the cumulation of probabilistic knowledge derived from the learning data. This way, it was possible to develop a new class of probabilistic models, which were also evaluated with success.
Currently, the group is participating in the TREC (Text Retrieval Conference) initiative, where indexing and retrieval with free text terms is performed for a large database with 2 GB of text. For the probabilistic indexing and retrieval methods developed on the basis of the general approach described above, the results from the 1st TREC conference showed good retrieval quality.
A second major research activity aims at the integration of probabilistic fact and text retrieval. It has been shown that probabilistic text indexing methods can be applied for fact retrieval involving vague criteria, that is, criteria which may be fulfilled with a certain probability for a specific database object; furthermore, imprecise or missing data can also be integrated within this framework. Based on this concept, the integration of text retrieval and vague fact retrieval has been proposed by combining a linear data model with a classical text retrieval model. As a further development, a first version of a probabilistic relational algebra suited for this integration has been developed. However, in this approach text and facts are still treated as separate attributes in the query as well as in the database objects, only the retrieval model is identical.
Norbert Fuhr
Norbert Fuhr (born 1956) holds a diploma and a doctorate degree in Computer Science from the Technical University of Darmstadt, Germany.
From 1980-87, he worked as scientific assistant and from 1987-91 as assistant professor at the Technical University of Darmstadt. Since 1991, he is professor in the computer science department of the University of Dortmund, Germany.
He is the current chair of the IR specialist group of the German computer society (GI) and treasurer of European IR specialist group CEPIS-IRSG. He has been member on Program Committees of major international IR and database conferences.
In research, he has worked on the development of an automatic indexing system for a controlled vocabulary and he developed and evaluated several probabilistic retrieval models. For materials databases, he performed a study on user-friendly design and appropriate retrieval mechanisms. In order to show that IR approaches are useful for non-textual databases, too, he developed a probabilistic retrieval model for vague queries and imprecise data in databases. Based on this work a model for the integration of IR and database systems was devised.
Recent Publications:
Probabilistic Information Retrieval as Combination of Abstraction, Inductive Learning and Probabilistic Assumptions. To appear in: ACM Transactions on Information Systems, 1993 (with U. Pfeifer).
A Probabilistic Relational Model for the Integration of IR and Databases. To appear in: Proceedings of SIGIR-93, 16th ACM Conference on Research and Development in Information Retrieval, 1993.
Integration of Probabilistic Fact and Text Retrieval. In: Belkin, N.; Ingwersen, P.; Pejtersen, M. (eds.): Proceedings of SIGIR-92, 15th ACM Conference on Research and Development in Information Retrieval, pages 211-222, 1992.
AIR/X - a Rule-Based Multistage Indexing System for Large Subject Fields. In: Proceedings of the RIAO'91, Barcelona, Spain, April 2-5, 1991, pages 606-623, 1991 (with S. Hartmann, G. Knorz, G. Lustig, M. Schwantner, and K. Tzeras).
A Probabilistic Framework for Vague Queries and Imprecise Information in Databases. In: McLeod, D.; Sacks-Davis, R.; Schek, H. (eds.): Proceedings of the 16th International Conference on Very Large Databases, pages 696-707. Morgan Kaufman, Los Altos, CA, 1990.