Information Technology Reference
In-Depth Information
The basis for understanding of the importance of this approach, as noted Arbib,
is that:
“The human brain is a calculator oriented on action. This implies that such
systems (of a man, animal or remote controlled robot) must correlate the action
itself and its result so that to build an internal «model» of a phenomenon.
-The human brain has a hierarchical layered organization. It is extremely
important that no single-level model is able to reproduce its functions.
-The human brain is not a system of the information processing with the
consequent execution of operations.”
In the late 1990s of the last century, Hans Moravec published a forecast for the
development of computer technology, which he called:
When the existing com-
puters will be equal to the human brain?
Based on the number of neurons in the
brain and the number of synapses (10
11
and 10
14
, respectively) and upon having
analyzed Man vs. Machine chess game, he concluded that the performance of a
digital computer (~10
8
MIPS) will be sufficient to simulate the functions of the
human brain. According to his estimates such computer systems should appear in
the first half of the twenty-first century, not earlier than 2020.
In this article, attention was drawn to three features of modern information
processing devices that are important for understanding the problem being
considered.
First of all, virtually any discussion of the perspectives of development of the
digital technology is related to solving intellectual problems (problems of artificial
intelligence). Indeed, this direction has acquired critical importance during the last
decades due to the need to understand and effectively control large dynamic
systems. As noted above, economic and social problems, traffic management,
global communication, weather prediction, assessment of environmental pollution,
and a host of other equally important aspects of modern society are among the most
pressing tasks. Virtually all of these tasks can be better handled by man than by
modern digital computer and can be attributed to problems of artificial intelligence.
Secondly, Moravec notes that some users of supercomputers, including
G. Kasparov, “feel” the rudiments of intelligence in computers. But this can rather
be attributed to self-deceptiveness. Even the designers of Deep Blue—the computer
which played with Kasparov—deny the presence of intelligence in the machine,
believing that it is “only a big database on openings and endgames, incorporated
and advanced functions, which are proposed by chess grandmasters, and, espe-
cially, the high efficiency, making it possible to examine the current situation on
14 moves onward.”
The third important observation is that a digital computer with the capabilities
approaching the human brain will represent a super complex system with tremen-
dous cost of creation and operation. Therefore, computers being created are funda-
mentally inferior to man in terms of efficiency, simplicity, and cost of solving
intellectual problems.
In general, solution of problems of high computational complexity, which
basically come down to the problems of artificial intelligence, leads to a manifold
increase in computational resources required to address them and forces developers
