uucp@ucbvax.UUCP (07/11/84)
From: decvax!mcvax!ariadne!simon (Simon Gibbs) The Institute of Computer Science in Heraklion is an active participant in ESPRIT (the European Strategic Programme of Research and Development in Information Technology). ESPRIT is a 5 year programme commencing after a 1 year pilot phase in fall 1984. The programme is comprised of projects of precompetitive R&D belonging to 5 areas: advanced microelectronics, software technology, advanced information processing, office systems, and computer integrated manufacturing. A project is undertaken by a group of partners which may include companies, universities, and research institutes. The total ESPRIT budget is ~1.5 billion ECU, in general 1/2 of a project's funding comes from the EEC and 1/2 from it's partners. ESPRIT has 12 main industrial partners who are expected to participate, they are: GE(UK) CGE(F) Telefunken(D) Olivetti(I) ICL(UK) Bull(F) Nixdorf(D) STET(I) Plessy(UK) Thomson(F) Siemens(D) Philips(NL) At the institute we are involved in 3 ESPRIT projects that were begun in the pilot phase and will most likely be continued for the full period. A proposal has also been made, and it seems to have a good chance for succeeding, for a fourth project to begin this fall. Each project is of the 50-100 man year size and extended over 5 years. A short summary of the projects, taken from their proposals, follows: ************************************************************ ESPRIT Pilot Project Workplan sub-area 4.4.1 Development of an Experimental Mixed-Mode Document Filing System This project will develop a prototype of a file server for office documents. The server will handle structured objects consisting of text, image, audio and attribute data (such objects are called "multimedia documents"). The operations will include document type specification (in which a particular structuring of components is identified), document storage (which requires mapping the components to physical storage), query processing (where the server receives a query from a client and sends back the objects satisfying the query) and operations on collections of docu- ments (such as creating or deleting a collection, retrieving a collection, and browsing through a collection). The server will exist in a network environment contain- ing user workstations, a high bandwidth communication facil- ity, and special devices such as printers, scanners, and OCR. Although the project is not directly aimed at these components, they will be considered in determining the func- tionality of the document server. The prototype server will be implemented in an incre- mental way: at the beginning of the project a first version will be implemented as quickly as possible, dealing with text and data only. This version will provide the basis for subsequent extension with image-handling and voice-handling functions. Partners Olivetti (I) Instituto Elaborazione Informazione (I) Cretan Research Centre, Institute of Computer Science (GR) Mnemonica (GR) ************************************************************ ESPRIT Pilot Project Workplan sub-area 4.2.1 Intelligent Workstations After a first phase, that lasted one year from Sep- tember 1983 to September 1984, the project on Intelligent Workstations will aim at the implementation of a complete system including office workstations and servers connected with a local network. The intelligence will reside in the workstation and the servers. Cooperation between all the components will take place in order to provide "Intelligent Assistance" to the users in their problem solving activi- ties. For this purpose the project is divided into five different modules and will run during five years. The first module is the assembly of the hardware pieces to construct the workstation. Dedicated hardware such as filtering mechanisms, that has been developed during the first year, will be taken into account when necessary in the workstation and/or servers. The hardware will be selected from products available on the market. The second module will be the operating system of the works- tation. It will be network oriented, insure multitasking and synchronisation with all the necessary mechanisms for increasing the reliability of the distributed system. The third module will be the implementation of the LISP environment and knowledge representation system that has been selected after the first year. The forth module will focus on the layer of office system tools that embed office semantics. The fifth module will deal with concrete applications to be implemented with the office system tools. The sixth module will deal with the user interface, it will be divided into three parts: natural language processing and an author system; enhancing the user interface with mul- timedia features (sound and image); defining and implement- ing a conceptual model of the user interface. The project will benefit from the very interesting techniques of Artificial Intelligence to design an office system that embeds on the one hand knowledge about office procedures and organisation, and, on the other hand, the necessary tools for a natural interface with the user (natural language and multimedia). There is no product today which provides all these features in the office arena. The project will fulfill partially the goal of recently advanced projects in the US and Japan (e.g., the 5G project). Partners Bull Transac (F) INRIA (F) Vrije Universiteit (B) Katholieke Univ. Nijmegen (NL) Cretan Research Centre, Institute of Computer Science (GR) OCE (NL) ************************************************************ ESPRIT Pilot Project Workplan sub-area 3.3 A Logic Oriented Approach to Knowledge and Data Bases Supporting Natural User Interaction This project aims at building a system allowing for a natural dialogue between users and data and knowledge bases. The system will consist of the following layers: - A user/machine interface in natural language (German and English) enriched with graphical facilities. - A human oriented high level representation formalism based on (an extension of) first order logic. - A machine oriented formalism also based on logic but intended for efficient reasoning. - A "machine" integrating the data and knowledge stored in the machine oriented formalism with the knowledge and data stored in existing tools (data bases, application specific tools). To achieve this goal, it is intended: - To develop a high level knowledge representation formalism which combines features of knowledge net approaches, logic approaches and advanced deductive mechanisms. - To compare different methodologies of natural language processing, to select the most appropriate one, to adapt it to the developed high level knowledge representation formal- ism and to enrich it with graphical facilities. - To develop a logic based representation allowing for effi- cient reasoning, to implement an execution mechanism based on this formalism, to integrate this formalism with existing tools (data bases, application specific tools) and to develop an optimiser transforming the human oriented representation into the machine oriented one. - To test the effectiveness of the selected methods and approaches by developing some applications. The considered application domains are the querying of data bases (a subset of the representation formalisms is expected to be suffi- cient), the evaluation of commercial aircraft design con- cepts, semantic data modelling and the use of robotic sys- tems in a production environment. Partners Belgian Institute of Management SA (B) Scicon Ltd (UK) Scientific Control Systems GmbH (D) Fraunhofer Gesellschaft (D) Cretan Research Centre, Institute of Computer Science (GR) Subcontractors Technical University of Munich (D) Cranfield Institute of Technology (UK) University of Hamburg (D) ************************************************************ ESPRIT Main Project Workplan sub-area 4.4.1 Development of an Optical Storage Based Retrieval System for Multimedia Documents This project aims to develop a server for the long term mass storage of multimedia documents. The server would be one component in an overall office information system and would be dedicated to document archival. The project will thus concentrate on developing a server with very large storage capacity and will, in particular, investigate the use of optical (i.e., laser) storage devices. We can describe the functionality of the server in terms of the objects handled and operations on these objects. Regarding the first we require that the server handle structured objects consisting of text, image, audio and attribute data - we call such objects multimedia docu- ments. Here text refers to arbitrary length character data, image to either raster images or a representation based on graphics primitives, audio to digitized (and possibly compressed) voice, and attribute data to traditional numeric or fixed-length character fields. Also we use "document" in the general sense which includes such things as letters, memos, forms, reports and so on. The operations provided by the server include document type specification, document archival, and query processing and retrieval. In addition many of the other devices in the overall office information system may provide services that directly or indirectly involve the archive server. For example a workstation can provide an interactive query for- mulation program, special input devices (such as OCR and facsimile) can generate documents for storage on the server, and gateways can transform documents received from external sources to the form understood by the archive server. Partners Philips Data Systems (NL) Truimph Adler A.G. (D) Battelle Institute e.V. (D) Cretan Research Centre, Institute of Computer Science (GR) ************************************************************ Regards from ariadne, Simon Gibbs