Cognitive architecture

A cognitive architecture is a artificial calculative process which tries to simulate the behavior of a cognitive system (generally that of human), or which acts intelligently under respect of a certain definition. The architecture term implies an approach which tries to model the internal properties of the cognitive system represented and not only the external behavior.

Categorization of cognitive architectures

The next sub-sections present several criteria to categorize cognitive architectures.

Level of specification

Certain cognitive architectures are specified with many details, whereas others not. ACT-R is an example of very detailed architecture; it has even a data-processing implementation.

Foundations on research in experimental psychology

Sometimes cognitive architectures try to reproduce with fidelity the cognitive behavior of human according to experimental results (storage capacity of work at the same time to treat a limited number of dimensions of one problem, lapse of memory, etc).

More than one hundred of models (modeling of knowledge for a task) were carried out with ACT-R such as the simulation of the human behavior to solve the problem of the tower of Hanoi. For many models, comparison were carried out between simulation and of human to make sure of the accuracy of simulations.

Some famous cognitive architectures

ACT-R, developed by John R. Anderson and his colleagues.
SOAR, developed under the supervision of Allen Newell with the University of Michigan.
Copycat, by Douglas Hofstadter and Melanie Mitchell with the University of Indiana.
DUAL, developed with New Bulgarian University and Boicho Kokinov.
Psi developed under the supervision of Dietrich Dörner with Otto-Friedrich University with Bamberg, Germany.

Applications

the intelligent Systems tutoriels

Random links: