The Fourth International Workshop on Artificial Intelligence in Economics and Management

Article excerpt

* The Fourth International Workshop on Artificial Intelligence in Economics and Management was held in Tel-Aviv, Israel, from 8 to 10 January 1996. This article discusses the Main themes presented at the workshop, including the need for multiple methods in any system designed to solve real-world problems, the differences in the effectiveness of AI versus classic analytic techniques, and the use of AI techniques to customize products.

The Fourth International Workshop on Artificial Intelligence in Economics and Management (AIEM4) was held in Tel-Aviv, Israel, from 8 to 10 January 1996, with participants from 13 countries. The main themes that emerged during the workshop were (1) the need to use multiple methods in any system designed to solve major realworld problems, (2) a continuing interest in comparing the effectiveness of Al solutions with classic analytic techniques, and (3) a growing use of AI techniques to customize products to suit individual consumers.

As a matter of course, almost every presentation at the workshop touched on Al techniques in one way or another. However, a group of papers at the workshop had AI techniques as their main focus. Two of these papers dealt with t e relative effectiveness of different methods and two with the development of new or existing techniques. P. P. M. Pompe and A. J. Feelders (both of University of Twente) compared the effectiveness of machine learning, neural networks, and statistical methods; they found that neural nets provided better results on their data than the other methods. M. Leshno and Y. Spector (both of the Hebrew University of Jerusalem) focused on training set size for neural nets and compared them to statistical methods with the same result; that is, neural nets provided better classification than the statistical methods. N. Levin and J. Zahavi (both of Tel-Aviv University) found that genetic algorithms performed even better than a linear programming model on their problem. Thus, their conclusion was that Al techniques might provide better results than rigid analytic methods for a variety of Two additional papers focused on the Prolog representation of business objects within the three-tier client-server architecture (D. G. Schwartz, Bar-Ilan University) and on the application of image theory to the design of agents and societies of agents (Schwartz and D. Te'eni [Bar-Ilan University]).

Service to customers in the financial area was another focus of the workshop. Lange et al. described a system for customizing investment instruments for individual customers instead of confronting them with a set of standard instruments. The paper by P. Lenca (Telecom Bretagne and Credit Mutuel de Bretagne) proposed a methodology for acquiring the knowledge of bank customer advisers for incorporation in decision support systems and knowledgebased systems.

Business applications represent a broad spectrum of issues dealt with in a variety of ways at the workshop. Y. Reich (Tel-Aviv University) proposed a representation for integrating quality-function deployment tools with AI methods. …

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.