Neural networks, modular neural network architecture, generalization

The identification of object and the processing of the other information are accomplished. The computational properties of this ?more-systems? design were explored by constructing simplifying connectionist models. The most important property of these systems is that they are capable of decomposing tasks into subtasks. That is, they can eliminate connections so as to dedicate different portions of the system to learn different tasks.

The identification of object and the processing of the other information are accomplished. The computational properties of this ?more-systems? design were explored by constructing simplifying connectionist models. The most important property of these systems is that they are capable of decomposing tasks into subtasks. That is, they can eliminate connections so as to dedicate different portions of the system to learn different tasks.