Neat Networks: A New Way of Looking at Machine Intelligence

摘要

While much research has been devoted to learning and machine intelligence, the field is still in its infancy. For example, a technology that will allow for the heuristic exploitation of information domain regularities to reduce the time required for knowledge acquisition while concomitantly resulting in an increase in the reliability of the acquired knowledge is still lacking. Furthermore, contemporary learning mechanisms such as neural architectures are inherently incapable of such performance. The objective of this paper is to present a new way of looking at learning and machine intelligence which has applicability in many fields such as in robotics (e.g., UAVs), intelligent agents, data fusion, and cooperative sensing. Neural networks cannot compute Modus Ponens, have NP-hard learning algorithms where one or more hidden layers are required, and are subject to the theoretical limitations imposed by Goedel's Incompleteness Theorem. We propose to construct a new type of AI, dubbed, "Neat Networks" for the intelligent fusion and transference of knowledge.