On a common structure of intelligence in biological and technical systems
Department 914, Research Institute of Aviation Technology 24 Petrovka Street, Moscow 103051Small-size Enterprise RTS-Machine 5-1-18 Obrucheva Street. Moscow 117421, USSR
This is a survey of the general structure of an electronic computer-implemented, operation-oriented system, designed by the authors, which uses artificial intelligence mechanisms and is intended for the control of technical objects that function both in predictable and random environments. The system is based on processing knowledge, which is stored in a hierarchically arranged Knowledge Bank, and program mechanisms for adapting to and interacting with the External and Internal Worlds. The system has distributed program mechanisms, which are designed with a constant structure. It is independent of the purpose and environment of the system operation and the specific features of the controlled object. None of the program mechanisms are concentrated in any progrcun module. They are distributed in many modules, and therefore there is no single module responsible for the execution of a particular external function. The system is structured into separate program modules by internal procedures. The conceptual organisation of the Knowledge Base presupposes that the framework is structured according to functional, semantic and tier indications, i.e. the structured description of knowledge, by the system, of the external environment and its possible behaviour in it. The possibility of multiple use of the same elements of the lower tiers of the Knowledge Base by highertier elements makes the proposed Knowledge Base very efficient. If the cerebrum is considered at a structural level, there appears to be an amazing similarity between the structure and the mechanisms of the above system and the structure of the cerebral cortex, as suggested previously by Edelman and Mountcastle. Each mechanism of the cerebral cotex structure has a structural analogue in the described system. Apparently, in a structural sense the described operationoriented system comes very close to the structure of intelligence in biological systems and could be used for its study and simulation.
Received on February 19, 1991; accepted on July 24, 1991