ARTIFICIAL INTELLIGENCE ALGORITHMS ON LINE FOR BALTIC COUNTRIES

Alexei Gvishiani

The four TELEBALT project countries (France, Latvia, Lithuania and Estonia) are among the leaders in algorithmical aspects of Artificail Intelligence (AI) studies. Thus, it makes it natural that these countries will join their efforts in this modern important direction of nowadays research. TELEBALT project plans to help to make first effort in this direction.

Computer science departments of Vilnius and Riga Universities, Institute of Cybernetics at Tallinn Technical University, Institut of Physics of the Earth in Paris and CODATA consider to start working together in the framework of TELEBALT project.. The goal of this collaboration is creation of joint on-line algorithmical resource for wide usage of research communities of the Baltic states and other European countries. This activity will be developed in close collaboration with the four TELEBALT partners (contractors) – EDNES (France), Open Latvia (Latvia), INFOBALT (Lithuania) and Inforing (Estonia). The first meeting on « TELEBALT Algorithmical Resource (TAR) » took place in Tallinn, Estonia on 5 July 2002. Prof. Jaan Penjam, director of Institute of Cybernetics at Tallinn Technical University, Dr ; Jaan Kalda, leading researcher in AI algorithms of this institute, Prof Jean Bonnin (University Louis-Pasteur, France) and Prof Alexei Gvishiani (Russian academy of sciences) discussed the first steps of the TAR planning. As the next step it was decided to organize a special meeting on TAR at the time of TELEBALT conference in Vilnius 22-23 October 2002.

The TELEBALT Algorithmical Resource (TAR) will mainly include three types of algoritms : vector classification with learning, clustering in finite sets and time series recognition. It is planned as on-line accessible collection of AI tools applicable to wide range of scientific and business oriented data. Well known fields for TAR algorithms application are geophysics and Earth science data in general, financial analysis and prediction and medical data recognition problems. TAR can be also useful in the problems of human resources distribution. In this way it can provide an important contribution to telematic applications in searching of new employment opportunities for the Baltic states. The latter is one of the main TELEBALT targets.

TAR is planned to be designed as a Web based resource efficiently accessible on line by the project participants and collaborating organizations from TELEBALT countries. It will be also made available on request to other IST projects interested to collaborate with TELEBALT on this matter. Each algorithm included in TAR will be represented by the following items : summary description, detailed mathematical description , references, the computer programme accessible remotly, conditions of the algorithm usage.

TELEBALT Algorithmical Resource: Artificial Intelligence Algorithms On-line

In the discussion have been participating: Dr. Alexander Beriozko (EDNES, France), Prof. Jean Bonnin (University Louis-Pasteur, France), Dr. Vytautas Cyras (Sintagma, Lithuania), Dr. Erinija Pranckeviciene (Kaunas University of Technology, Lithuania),  Prof. Laimutis Telksnys (Institute of Mathematics and Informatics, Lithuania). Although there exists already libraries of AI tools on the internet, all participants of the discussion have agreed that the need for such library in Baltic countries is obvious. As a result the following suggestions were made:

1. To set up a WWW page in the INFOBALT web site for disseminating information concerning TAR( Telebalt algorithmic resources) and collecting feedback suggestions from the Baltic countries and all interested parties.
2. To find out what kind of AI resources are available in Baltic countries and countries participants of the TELEBALT project. The findings are going to be presented in the next TELEBALT event in Riga 2003.
3. All interested to contribute to the library their algorithms are encouraged to provide a summary of the resources they consider to share through the INFOBALT web site.
4. The main stress of the library should be on the reports of the usage of the algorithms in the domain specific areas. The intention is to provide a research community and all potential uers with the positive experiences and drawbacks on using some particular algorithm in some particular problem.
5. To consider including the domain specific data for the testing of TAR. The vision is that later this site can be developed into the benchmarking site attracting an attention of the wider audience.
6. On the basis of the collected summaries and the resources provided by the contributor’s an actual proposal for the TAR project as one of the successors of the TELEBALT project is going to be presented.

Chairwoman of the discussion
Dr. Erinija Pranckeviciene

Presentation

The Artificial Intelligence section of LIKS