In principle, there are two ways to overcome bottlenecks that hinder the development of intelligent technical systems: to elaborate new and more effective methods of knowledge representation and manipulation, and co-operation between separate parallel systems. The aim of this project is to promote both these trends and develop for solving practical engineering problems an artificial intelligence environment based on high performance computer networks.
In the scope of this research automatic deduction methods and their usage in constructing computer programs are investigated. As the result the artificial intelligence system NUT has been developed.
The NUT system is a knowledge-based programming environment that combines object-oriented (procedural, modular) and declarative (structural synthesis of programs) programming paradigms.
THE NUT SYSTEM ALLOWS
- specification of problems from different application areas as packages of knowledge containing classes, objects, programs, pictures, texts etc.;
- integration of different models of computations (program synthesis or message sending for manipulation objects and classes);
- creation of interactive systems using binding of graphical representation with objects.
The NUT environment includes the language interface for C allowing to use programs written in C as components of the NUT system.
THE NUT ENVIRONMENT CONTAINS
- windows-based interactive user interface;
- language processor including program synthesizer from formal specifications of external behavior of the programs (can be regarded as proof search in intuitionistic propositional logic);
- graphics facilities including Graphics Editor and the Scheme Editor (a tool for visual programming).
The NUT system is suitable for fast prototyping in different problem areas such as CAD/CAM, control systems, simulating systems, data modelling etc.
The results have been applied in modelling radar coverage of Estonia (V. Kotkas, R. Lõugas, M. Harf, J. Penjam) and in automating software engineering systems (A. Kalja, K. Kindel, T. Tiidemann (TTU)). Applications in higher education are associated with carrying out Tempus-projects (J. Penjam, M. Perkmann, J. Vain) and elaborating artificial intelligence related courses (E. Tõugu, J. Penjam, A. Kalja).
PARTNERS
The NUT system is a cooperative development of the
Institute of Cybernetics at Tallinn University of Technology (kernel, language processor) and
Software Engineering group (
http://www.it.kth.se/labs/se/nut.html) of the Teleinformatics Department of the Royal Institute of Technology, Sweden.
The NUT group can be reached by e-mail nut@it.kth.se