A Computational Model for Multiple Goals

A hierarchical attention-based neural network architecture, based on human brain guidance, for perception, conceptualisation, action and reasoning

“…However one can easily transform a list of serially or concurrently executing goals to a suitable neural network structure, as explained in [38]. Thus the representation and the execution of a Plan can take place by using solely sub-symbolic (neural network) means, as are the rest of the representations used in GNOSYS.…”

“…More complex actions can be represented by a list of primitives, i.e. a plan of primitive sub-goals (for more details on the second aspect of a goal see [37,38]). …”

“…This request comes from two primary sources: a user request or a request from the motivational system. The realisation of the request takes the form of a Goal object, which is characterized by a number of parameters (see [38] for details). The Goal object represents either a complex high-level goal or a primitive one.…”