Community Research and Development Information Service - CORDIS

Abstract

In this paper two attempts to use artificial intelligence (A.I.) techniques in the framework of nuclear reactor reliability data bases (namely, the European Reliability Data System - ERDS) are described. The first attempt concerns the design of an intelligent interface aimed at supporting end-users in accessing in a friendly, correct and effective way the ERDS, which is a complex system consisting of three data bases loosely coupled. The interface includes both natural language understanding and expert capabilities; furthermore it features modular architecture meeting the needs of technical and non-technical users. The requirements of ERDS users are investigated and the design of an intelligent interface is proposed. The characteristics of the natural language query interface and the expert interface are later discussed and the present state of the project is given. The second attempt is devoted to the development of a generalized semi-automatic transcoding system based on expert system techniques. This system is applied to the same ERDS which receives data from various national data systems. In the paper, the overall architecture of the system is presented and its main features are outlined. The formalism for representing specific knowledge (input and output models, transcoding rules and algorithms) is then illustrated. The project places special emphasis on a clear and practical interface for the user during the transcoding process and for the management of the transcoding rules.

Additional information

Authors: COPOBIANCHI S, JRC - ISPRA ESTAB. (ITALY);CARLESSO S, JRC - ISPRA ESTAB. (ITALY);BARBAS T, JRC - ISPRA ESTAB. (ITALY);KOLETSOS A JRC - ISPRA ESTAB. (ITALY), JRC - ISPRA ESTAB. (ITALY)
Bibliographic Reference: IJCAI'87 - INTERNATIONAL JOINT CONFERENCE ON ARTIFICIAL INTELLIGENCE, MILANO (ITALY), AUG. 23-28, 1987 WRITE TO CEC LUXEMBOURG, DG XIII/A2, POB 1907 MENTIONING PAPER E 33554 ORA
Availability: Can be ordered online
Record Number: 1989126037801 / Last updated on: 1989-01-01
Category: PUBLICATION
Available languages: en