Artificial self-awareness:
- is the idea of a variety of "Artificial Intelligence."
This idea may be interesting, but it is stated amateurishly - I have 20 years left the topic, and the current state of this issue has not been studied.
I decided to put the idea in a raw form, from memory, "to the lover." - If this direction has long been developed, then you can not read further. - I judge the state of the issue not from studying modern theses, but only on a superficial user observation.
I once studied scientific psychology. I will state here the course of my thoughts - at that time:
"The psychology of self-awareness":
The basis of the human intellect is self-awareness.
Similarly, the most promising direction of "Artificial Intelligence" should be "Artificial self-awareness."
One of the reasons why this direction does not develop is the lack of science "Psychology of Self-Consciousness."
Used in psychology, the term "self-awareness" does not mean realizing the self-awareness, not separating the "I" from the rest of the psyche.
If we again draw an analogy with the human intellect, the existing directions of "Artificial Intelligence" are not reflexive, but reflexes.
Another non-optimal direction of "Artificial Intelligence" goes "from below" - from the interactions of neurons in the network. - The physiology of the brain in humans and mice is the same.
The basis of human intelligence is the mechanism of self-awareness.
Auto-programming:
I saw the concrete realization of "Artificial self-consciousness" in auto-programming.
That is, in the creation of programs that would write / adjust the program code - based on the results of the analysis of the activity of this code, and using knowledge of the domain.
Objective and similar directions of development of programming languages assume, one way or another, human participation.
In 1996-97 I experimented with programs writing code.
But, I stepped back - the problem was too complicated and large.
In addition, I was not available for similar studies.
And I also had a paranoid thought that in Huxley's society Artificial self-awareness can be dangerous.
I realized that I could not do anything sensible myself, abandoned the programming, and decided to reorient myself to finding unanimous people who would be interested in the ideas themselves, and with whom projects could be realized together.
Labor productivity:
The program was supposed to find and build code for a very briefly set goal / task, which would achieve this goal.
The idea was to multiply the productivity of labor. - Writing the code was shifted to the program.
You can say I tried to create a "translator" of an even higher level.
Programming languages "high-level" replace the generalized commands with whole "sentences" in low-level machine language.
But I did not try to make even more generalized teams. - This direction has already reached saturation. - Depending on the goal / task, the teams had to be modified - the programmer. The complexity of programming grew, but not the productivity of labor.
The functional grew, but, as in the USSR, due to the increase in the cost of programming.
I also tried to teach the program to independently search for the realization of a human goal / task.
It would be nice if the program asked for ambiguities, replenishing its knowledge.
But before that I did not get there. - I limited myself to the formulation of the problem again using syntactic rules.
Solution search:
The most primitive approach to finding solutions is a search by rules. -
- The program tries numerous code options.
But it builds the code according to the rules to make it workable.
There are numerous theories, for example, the search for a move in the game of chess, which by many orders reduce the search of options, based on the analysis of the subject and the effectiveness of one's own activity, with the use and accumulation of knowledge.
Areas of use:
We need to start with the search for narrow special areas, where increasing the productivity of labor will have the greatest economic effect.
They can be online-services in different spheres.
Search engine information and help:
Statement of the problem and the accumulation of knowledge here can be an assessment of users and statistics of the behavior of different people.
Where human control is difficult, for example, in asteroids.
|