New opportunity to create artificial intelligence

One of the many popular definitions of AI on the Internet (from Wikipedia) is as follows: “Artificial intelligence is an experimental scientific discipline whose task is to recreate intelligent reasoning and actions using artificial devices,” but opinions were seriously divided into “strong” and “weak »AI:

• -    the term "strong AI" was introduced by John Searle, this is when the AI will be a mind with emotions and versatility, like a human mind;

• -    supporters of "weak AI" consider AI to be a tool to help a person in his activities. And a lot has been achieved here.

It should be noted that R. Penrose (English physicist and mathematician) in his book "The New Mind of the King" actually spoke about the impossibility of "strong AI" in the near future, due to the existing level of science capabilities. However, at the same time he agreed that certain achievements in this direction are possible. Such achievements, for example, include voice perception and text printing, respectively, its translation into foreign languages.

But with the possibilities of making various kinds of decisions, the scientific world cannot come to a consensus, at present there are more than 500 different methods, mainly of an indefinite psychological nature.

In addition, one of the important and difficult problems in decisionmaking is rightfully considered the need to overcome the so-called “psychological inertia” (PI), when the available knowledge prevents an objective consideration of the problem situation.

Relevance. Understanding the importance of work on AI and taking into account the above, we can conclude that any significant and effective contribution to the creation of AI should be considered relevant and timely, especially when it comes to formalized machine decision making of a high intellectual level.

Purpose, task, justification, method. The purpose of this work and the article is to show the possibility of machine-made deterministic decision-making for almost any set tasks using information technology.

The task that was solved at the same time was to simulate the human mental process by machine methods. It was based on my many years of experience as a designer at the Russian Federal Nuclear Center VNIIEF and as a technical director at a private company for the development and production of import-substituting polypropylene steam-hydro-insulation. Especially, given the fact that working according to a formalized method of decision-making, about fifteen inventions were patented in areas of knowledge where it seemed that everything had been invented long ago.

The way of making decisions was considered step by step: from setting tasks to their final identification into a real object. As a result of this consideration, in 2017, a patent for invention No. 2652501 "Module for searching a block of information by input data" was obtained.

As in any invention, this one contains the most essential features and their combinations, which make it possible to achieve the set goal:

introduction to information processing of search for functional factorial sorting of input data. Due to this, at the beginning of the formulation of the problematic problem, the module is forced to explore all possible significant input data, functionally sorting and grouping them;

an additional level of features characterizing details has been introduced, starting with the finally selected input data, which in turn allows the module to carefully - more accurately prepare the information field for the further work of the module. This is what ensures the reliability of the module: an additional level of details has been introduced about the features of the input data in the form of their characteristics and significance;

the accuracy of the result - the required block of information is increased by the introduction, in comparison with known other similar devices, additional blocks and additional inputs, and outputs between them, making it possible for the module to repeatedly specify the characteristics of the search object;

due to the "translation" of the decision-making process from the psychological to the information-machine one, it is difficult even to imagine that the machine will need to overcome the PI, since it will work with formalized information data, features and characteristics.

The problem remains, to convince the scientific community of the effectiveness of the described approach in making decisions determined by carefully worked out input data. This article, in particular, has such a purpose.

Novelty. Due to the fact that the material presented was recently patented, the world novelty has been confirmed.

Conclusion and conclusion. The possibility of machine decision-making based on input data is shown.

This allows serious progress in the creation of AI in its most difficult and responsible area - decision making.

Moreover, the algorithmic principle of the module will be very useful in ordinary project work, as well as in making managerial and organizational decisions.