Agents Detecting User´s Intention.

ARMENTANO, MARCELO G.; AMANDI, ANALIA A.

Buenos Aires

Simposio; Simposio Argentino de Inteligencia Artificial - 32° Jornadas Argentinas de Informática e Investigación Operativa; 2003

SADIO, Sociedad Argentina de Informática

As computers are used for more tasks and become integrated with more services, users will need help dealing with the information and work overload. Interface agents radically change the style of human-computer interaction. The user delegates a range of tasks to personalized agents that can act on the user´s behalf. In this way, the agent gradually learns how to better assist the user. This paper presents an approach that shows how an interface agent can detect intervention situation to cooperate or advise the user by observing user’s activities on an application. The need of understanding what the user is doing lies in the fact that the agent has to act always in the context of the user intention. Acting in this way, the agent would not bother the user in an improper moment and the user can always feel that it has the control of the application. The approach presented is a portion of an own software architecture, named AGUSINA, which defines specific components to manage Agent-User interactions.