Understanding Capacity as the Principle Difficulty in Building Artificial Intellect

Abstract

This paper discusses those factors by the influence of which the great task of building artificial intellect had been forgotten para­doxically just in the modern computerized life. The dream of building thinking machines got almost practical sicnificance in the mead 20^th century when were the first elec­tronic computing mac?ines and took place the short period of formation and intensive development of cyber­netic science – one of the 20^th century most significant achievements. But it was realized soon that the optimistic expectations of building thin­king machines in a decade or two were not rea­lized. The author of the paper mentions two main difficulties in building artificial intellect: the prob­lem of developing the general algorithm of prob­lem solving and the problem of computer under­standing of natural language textual information.

The author claims that he has succeded de­veloping 40-step general scheme of problem solving that he conventionally calls Problem Solving General Algorithm. He is convinced that to develop a complete general algorithm of problem solving one has just to continue making more detailed the shceme he has suggested. The main steps of this general scheme of problem solving are presented in this paper.

In regard of the task of the problem of com­puter understanding of natural language infor­mation the author suggests, first, applying nume­rical indexes to different meanings of natural language words and phrases, second, introducing indicators for contexts of natural language texts. The paper suggests also a system of classification of the levels of understanding.