About emotions, programs and artificial intelligence
It so happened that last year we were engaged in teaching the iPhone to understand emotions. Not recognizing facial expressions and gestures is only one of the external manifestations of a person’s emotional state, namely, to understand a person’s state by his answers. The software product turned out to be more of a sophisticated and beautiful psychological test than computer technology, but it made our team (professional psychologists) deeply understand the topic of emotions.
If you believe what habrovchane write, any program, this is an algorithm - that is, a specific procedure that has causal logic. If you still simplify - then this is a set of "if-something" and computational actions between them. The main flaw in any program is the fundamental absence of the answer to the question “why?” Within the algorithm. The “if-then” logic is the logic “because”, but not the logic “why”. The meaning of any program is outside of it and is laid by the programmer at the time of creation. Hence, at the level of mind games, it follows that the program always experiences the same emotion - the emotion of self-realization. That is, all that the program needs is to execute itself. In this case, no complication of the program itself does not change the essence of the postulate. The program can behave differently, it can beat Kasparov in chess or serve you at McDonalds - but the only thing that she will “want” for, what will be the point of her existence is to fulfill herself.
If we look at the definition of emotions, then they determine the estimated attitude to the world. Emotions, as an element of the motivational-volitional sphere of a person, in themselves carry the answer to the question “why” Actually, they are this very "why." Look at the activities of people - people act in order to feel. Feel joy for children, pride of victories, anger towards enemies, interest in work, etc. Not because they feel, but in order to feel. Rational-logical sphere of a person is built as an adaptation mechanism that helps more, more often and more to experience feelings.
')
Thus, it turns out that to program emotions is to program the meaning of the existence of a program. And we have already agreed that at the level of the algorithm, inside it, it is impossible to answer the question "why."
Hence, it is obvious that the problem of programming emotions is not solved in principle.
What does a simplified model of emotions look like? In general, it is not difficult, it is enough to have:
1. Subject
2. The external environment
3. The internal algorithm for assessing the "value" of environmental factors for the subject
The subject perceives the external environment through the prism of its “value” for him and modifies his behavior depending on the result. Someone may declare that, for example, Deep Blue (subject) was happy to strengthen his chess position and grieved if he was to lose a pawn (value estimate). But it is not. It didn’t matter to the program whether it won or not, and winning or losing did not have any value for it (the program).
A critical factor is the external environment and the subject. And if the external environment is present, thank God, then, in order to feel the emotions, the program will first have to become a subject, that is, become alive and acquire its own, independent value. The order of creation of artificial intelligence does not differ from the order of appearance of ordinary intelligence. First you need to become alive, then learn to feel, and only after that you become rational.
If you believe what habrovchane write, any program, this is an algorithm - that is, a specific procedure that has causal logic. If you still simplify - then this is a set of "if-something" and computational actions between them. The main flaw in any program is the fundamental absence of the answer to the question “why?” Within the algorithm. The “if-then” logic is the logic “because”, but not the logic “why”. The meaning of any program is outside of it and is laid by the programmer at the time of creation. Hence, at the level of mind games, it follows that the program always experiences the same emotion - the emotion of self-realization. That is, all that the program needs is to execute itself. In this case, no complication of the program itself does not change the essence of the postulate. The program can behave differently, it can beat Kasparov in chess or serve you at McDonalds - but the only thing that she will “want” for, what will be the point of her existence is to fulfill herself.
If we look at the definition of emotions, then they determine the estimated attitude to the world. Emotions, as an element of the motivational-volitional sphere of a person, in themselves carry the answer to the question “why” Actually, they are this very "why." Look at the activities of people - people act in order to feel. Feel joy for children, pride of victories, anger towards enemies, interest in work, etc. Not because they feel, but in order to feel. Rational-logical sphere of a person is built as an adaptation mechanism that helps more, more often and more to experience feelings.
')
Thus, it turns out that to program emotions is to program the meaning of the existence of a program. And we have already agreed that at the level of the algorithm, inside it, it is impossible to answer the question "why."
Hence, it is obvious that the problem of programming emotions is not solved in principle.
What does a simplified model of emotions look like? In general, it is not difficult, it is enough to have:
1. Subject
2. The external environment
3. The internal algorithm for assessing the "value" of environmental factors for the subject
The subject perceives the external environment through the prism of its “value” for him and modifies his behavior depending on the result. Someone may declare that, for example, Deep Blue (subject) was happy to strengthen his chess position and grieved if he was to lose a pawn (value estimate). But it is not. It didn’t matter to the program whether it won or not, and winning or losing did not have any value for it (the program).
A critical factor is the external environment and the subject. And if the external environment is present, thank God, then, in order to feel the emotions, the program will first have to become a subject, that is, become alive and acquire its own, independent value. The order of creation of artificial intelligence does not differ from the order of appearance of ordinary intelligence. First you need to become alive, then learn to feel, and only after that you become rational.
Source: https://habr.com/ru/post/146900/
All Articles