To process an external environment and an internal environment in an integrated manner as a cognitive behavior system so that the external environment and the internal environment can be recognized as a symbolic representation.
The present invention is a recursive neural network having a plurality of I/O parts composed by combining standard modules of consciousness, and consists of two information circulation circuits c1 and c2 (feedback circuits) which synchronize and excite artificial nerve cell groups called neurons, and a simple input part (a). The I/O part on one side of the circuits is also a symbol representation part for achieving various kinds of behavior and cognition of the module. The module conducts learning so that the information of the input part (a) and the information of the output part b always match, and, as a result, learning of various functions is possible. The consciousness system is composed by linking the modules at the respective I/O parts, and implements cognition and behavior by the respective modules exchanging symbolic representations using the respective I/O parts.
COPYRIGHT: (C)2007,JPO&INPIT
| JP6309293A |