Abstract

A user is a system capable of creating and pursuing individual goals. Is it possible to design and implement an artificial user? Traditional artificial systems focus on how achieving a given goal. Most learning algorithms look for an optimal solution of a problem, given a set of optimization criteria and a goal (or a set of goals). However, real agents and real users have to develop new goals in order to cope with their environment. They must find “what” they want to achieve and not only “how”. The development of completely new goals on the basis of the interaction with the environment is here defined the “what” problem. In this paper we will try to define it and we will propose an architecture capable of addressing it. Such an architecture is proposed as the foundation for an artificial user.