Artificial Cognitive Neural System Based on a Novel Neuron Structure and a211 Reentrant Modular Architecture with Implications to Machine Consciousness

摘要

The author proposes that artificial cognition, as opposed to rule-based211u001eartificial intelligence, should emulate human cognition. As a starting point, 211u001ecognitive functions and processes are reviewed and discussed. The following 211u001eelements are identified: Distributed signal representation with meaning and 211u001esignificance, perception process, attention, associative learning and recall, 211u001ematch/mismatch/novelty detection, pain/pleasure as rewarding and motivating 211u001efunctions, processing by inner representations, the faculty of introspection. A 211u001enovel non-numeric associative neuron, suitable for distributed signal 211u001erepresentation and associative learning is proposed as the basic processing unit. 211u001eA cognitive neural system that processes information by inner representations, 211u001einner imagery and inner words is presented. The operation of the neuron and the 211u001ecognitive system is verified by computer simulations incorporating visual and 211u001elinguistic modalities.