Center for Strategic Assessment and forecasts

Autonomous non-profit organization

Home / Science and Society / New in Science / Articles
Evolution of intelligence: why robots emotions
Material posted: Publication date: 19-08-2019
Emotions and intelligence, physics and poetry. How much longer the opposition of those categories?
It would seem, we all know that emotions interfere with the intellect and we appreciate the cool, admire their ability not to give in to emotions and act rationally. On the other hand, the lack of emotion too, we do not really like. It's possible that not everyone likes pedants and crackers and when they Express emotions we sometimes think that it is the very humanity.

What are emotions? Exclusive whether it is the quality of the person or them still possess animals? And finally, do you need emotions to robots and whether they have them at all?

All who are interested in such matters and likes to philosophize, welcome under kat.

Introduction

This article is a continuation of the previously published under the title
Evolution of intelligence: the beginning. It outlines a pretty simple idea, what if we assume that intelligence is not appeared at once, and went the evolutionary path from very simple forms to modern models of the latest generation (of human intelligence), the traditional definition of intelligence, sharpened by human will clearly require revision in the direction of greater universality. In addition, the article for the convenience of discussion introduced the classification of the evolutionary levels of intelligence, as some analogues of the generation equipment.

A new definition

Try more widely to define intelligence.

Intelligence is the observed ability to solve the tasks set for it by the media

Because of its versatility it helps me to break away from conventional ideas about intelligence, as an exclusive ability of man and to look at the world with a less anthropocentric positions.

In addition, the definition focuses on the need to monitor this ability. We do not yet know how to measure the potential of intelligence. In works of fiction there is often a device that issues from the figure of the intellectual potential. However, in the real world, to measure some aspect of intelligence we use exams and tests, including, in fact, from the individual task and during the test to observe and measure the scores that the ability to address them.

Individualization

The main characteristics of the evolutionary level of intelligence (as analog of generation in the technique) was chosen personalisation. Judging by the comments to the first part of the article this choice has caused confusion.

But we all are familiar with the individualization. When we choose clothes, so not like everyone else, baubles decorate yourself or your car, make a cozy peculiarity in the decoration of their standardized houses — all this is a manifestation of our individuality, manifestation of freedom of our inner world.In sum, many individuals give wealth. The wealth of opinions, ideas, decisions, points of view, expression and, ultimately, the intellectual wealth we, as a people and as a whole species. Conversely, the process of unification of people, the desire to "cut under one comb", "to measure one size fits all" — the modern synonymous of brainwashing, of belointeractive.

But whether to be unique or to unify myself, to be myself or be like everyone in our time is still our personal choice. At least, we are not limited in this constructive and our intellect, the human intellect, in its capacity offers us an infinite individual variety.

Technically, at the other end of the scale of individualization are algorithms. Different computers with the same program will always work the same way. And, precisely because of its form factor, the intelligence of such class has zero individual variety or zero level of customization.

Intelligence on the road to mind

Try to understand what are the evolutionary milestones are between zero and infinite individualization?

The first step — where each individual is a carrier of intelligence solves the problem in some different way, but the way these decisions [technically] unchanged during the life of the individual. If you add inheritance and selection, you get a quite clear the mechanism of evolution, when unsuccessful and ineffective solutions will be screened, and the successful and effective — to gain a foothold. In nature this is realized through unconditional reflexes. Let's call it intelligence I level of individualization.

The second step is when the individual is a carrier of intelligence has the [technical] ability to modify solutions to problems. Now he is the individual in the course of his life can choose a good solutions and to dismiss ineffective. In nature this is realized through conditional reflexes. This is the intelligence of the second level of individualization.

Step three — when the individual is a carrier of intelligence has the [technical] ability for concerted collective action. Coordination requires the development of conventional communication. The communication channel can be quite different: the gestures and sounds and smells and visual signals. The fundamental characteristic here is conventionality, i.e. the values of the signals are not rigidly predefined algorithms, and is the result of the agreement, a local group of individuals. An important bonus of this level — the potential ability for interspecies communication. It is already intelligence III level of individualization.

The fourth step is the appearance of abstract-logical thinking. As you know, the thinking is based on their ability to communicate, as abstract and logically we think in words, that is, communication units, which are semantic. It is the intelligence of the IV level of individualization or human intelligence. This level of intelligence is traditionally called the mind, and the man himself as a species — sapiens.

As we can see, each step is fundamentally expands the degree of individualization, gradually changing the potential values of this feature range from zero to infinity.

Robopsychology as it is

Due to the proposed higher classification of the level of intelligence we can more clearly understand how the improvement of robots, and their consequences.

If the evolution of intelligence is the path from the moment of individualization, it means the inevitable emergence of robots with individual behavior, so inevitably the question arises how this behavior is to control and regulate.

A robot may not injure a human being or, through inaction, allow a human being to come harm. (C) "3 laws of robotics", A. Azimov, 1942

As we see from the birth dates of the cited works, the issue of control of the behavior of the robots is interesting to mankind for the first decade. In this case, a recognized classic of science fiction went the traditional route of law and himself in their works have shown the futility of this option due to the ease of change of its interpretations.

The law that pole, where turned, back and left (C) popular wisdom

But if the option is constrained by the legal language of the hopeless, how to solve the problem of control and regulation of the behavior of the robot? Let's look at how this challenge of nature. Since II is the individualization of living beings has received a technical possibility of some freedom of their conduct. And what in this case with the interests of their own security, survival and prosperity of their species?

Nature has solved this problem in a very interesting way. Hardcoded remained the strategic objectives of the individual, and the tactical was handed over to the mercy of changeable individual behavior. The mediator between strategy and tactics become emotions. Every living individual is the level of customization was the ability to experience positive emotions when his behavior is in line with the strategic objectives and negative when contrary. Moreover, emotions can appear after the fact, after the decision, for example, as the joy and jubilation from the successful rescue in a moment of danger, and to precede and motivate the decision, for example, anxiety during the rut.

Love dries a man. The bull lows with desire. The rooster finds a place. The Marshal of the nobility loses his appetite. (C) 12 Chairs. Ilf and Petrov.

You can transfer this solution to the area of robotics and in a similar way to control the behaviour of robots by the promptings of the emotions to action and emotional reward for right action and punishment for wrong? Probably Yes. How? This will be discussed in the sequel.

It is possible that if the robot to experience emotion, its behavior we will be much closer and clearer and we can more easily enable robots to the community people.

Thanks and invitation

The author thanks Professor N. In. Khamitova for their invaluable help during the development of this theory.

The author invites everyone, and especially evolutionary biologists to participate in the discussion, and perhaps, fascinated, and join the work on the theory.


Source: https://habr.com/ru/post/464097/


RELATED MATERIALS: Science and Society