Dynamics in transition mental activity

The scene is set now for computer modelers to rake over and highlight the shape and accessibility of the reactive mental sites. Until recently, attempts to model artificial agents had to rake account of every functional unit in every architecture. The dynamics of mental states events is crucial to design agents because we need to rethink the role of mental states when tuning behaviours. Tiny changes to the shapes of these states make a big difference to what agents do, and only by inspecting how states interact we may find the most suitable structures and mechanisms for the mental activities. The paper presents how dynamics in transition mental states to agents' behaviour results from a set of properties and laws governing combinations of mental states. These agents can have mental states (beliefs, desires, intentions, expectations) and the activities of these mental states occur in the context of an agent architecture. This architecture, called SEM (Society of Mental States), is a nest of four sub-agents built around beliefs, desires, intentions and expectations. The characteristics of such agents are defined from the beginning as a basic set of desires, expectations, actions and beliefs, which is also made of strategies to interact with other agents and the world, as well. The interactions of such agents are the result of their mental states' dynamics and their actions, as we show in the paper. Namely, in order to have a pair of agents with tutor and learner behaviours it is only necessary to specify their mental states.