Wisdom key: On the right and wrong of Turing test

The Turing test is a method used by computer scientists and philosophers Turing in the last century to test whether machines (computers) are intelligent: when most people are unable to discern that their chat object is a machine, then we can think of the machine as having high intelligence.

This involves the question of "What is Wisdom": we first define wisdom as "the logical process of the life body's corresponding feedback to external stimuli". The basis for determining whether an object is intelligent is to "see if it is able to respond appropriately to the changes and stimuli of all external circumstances".

We all know that machines can only handle information in accordance with pre-set logic, and they obviously cannot handle situations where there is no "out-of-the-box solution." This is the principle of Turing's test: People can judge whether the other is a machine or a real person by using a dialogue to test if the other person is able to make appropriate feedback on any situation.

But Turing Test has a more obvious "point of contention", that is, the "feedback is appropriate" process is a "too subjective can not be objective quantification" process:

For example, when I ask, "Why can't the speed of light be surpassed?" "When, if the other person answers," because in the theory of relativity. (in a scientific way to answer the question head-on) "or" I don't quite understand, so why don't you just ask the mother (admitting that you're not on the subject). It can be said that these two kinds of feedback can be accepted by most people, if you can not accept any one of the feedback, it can only show that "it is not appropriate for you to feedback," and does not mean that "it is not appropriate for all people feedback."

This example can prove the "limitations of Turing Test" from the side: even if a machine or software is able to pass the Turing test, but because we do not have an objective "quantitative assessment" of the test process, the process of Turing testing can only illustrate an objective fact that "the feedback of this machine is appropriate for that group of people", If a different group of people to test the machine, it is likely to get a different result.

To look at the problem from another perspective, objectively speaking, if a machine wants to achieve the goal of "making the right feedback for any situation," it really needs to be "fully autonomous", because "we can't preset the solution for the machine", which means that the machine must be done autonomously. The process of identifying user intent and building a solution (instead of using a "matching preset solution" approach to the problem).

Or we can understand the correctness of the Turing test from this perspective: in order for the machine to be able to cope with all kinds of situations (intelligence), it really needs to master the human language system, because all human knowledge systems can be expressed in words, and mastering the language system means mastering all human knowledge, This allows the machine to make appropriate responses to any questions that humans ask.

But using dialogue to test whether the machine can properly respond to all situations relies too much on subjective judgment, this test process cannot objectively determine whether the process of the machine completes the conversation with a matching solution or whether it truly identifies the user intent and builds the solution autonomously.

Analysis here, some people may be in the heart of the confusion: how to test the machine is intelligent? Or is it impossible to measure intelligence objectively? What is wisdom?

When a computer encounters a "No preset solution", it feeds back a "system error please return" message, which we take for granted is inappropriate. However, although human beings belong to the "higher intelligence" species, can think can speak, but also can create and feel happy, but people will still encounter a variety of "can't find a good solution" situation, the person in this situation will be feedback "Sorry I do not". What happens if I return the computer feedback "system error please" instead of "Sorry I don't"?

Then there is a paradox: if people define wisdom as having to return the right conclusions about everything, then human "Sorry I won't" should not be counted as an intelligent response (only appropriate feedback). If people can accept the fact that wisdom is limited, it should be regarded as a kind of intelligent reaction that the computer's "system error please return".

People take it for granted that "artificial intelligence should never be made wrong", and people take it for granted that "wisdom is something that is out of the logical category." The idea is wrong, the human brain is limited by the human brain, computers are limited by computers, and the process of computer problem-solving is very similar to the process of logical thinking in the brain (in reality):

When you said to me, "It's a nice day." I almost don't need a logical thinking process to get a solution that says, "How should I respond to the other person's willingness to greet you?" because I have a "fixed feedback" experience in my mind that is similar to the process of a computer matching a preset solution.

If this is the first time I've ever been in a situation like this, I would first consider the true meaning of the phrase, "It's a nice day." For you and I are not very meaningful things, even if the premise of today or yesterday is still not very meaningful, then I can judge you say to me that the true meaning of this sentence is actually to me. " Polite greeting. "

Then I will consider how to respond to the "Other people take the initiative to greet you" situation, my experience tells me that although it is a kind of "goodwill behavior", but I do not need to make a "Thanksgiving Shong" too much response, so I am in "All feasible feedback plan" Pick out one of the most in line with the current situation (do not want to talk more on the way to work), and finally decided to reply to you a "Yes ah, you look very good."

If the computer to face such a situation, computers may only be built in the "weather" and "good" corresponding meaning, and then the computer needs to be based on the "grammatical logic" to combine these two words, and according to the "logical algorithm" to generate a conclusion. And so on until it reaches a final feedback conclusion. Whether or not it comes to this conclusion is correct, mainly to see whether its "logical algorithm" is strong enough.

We can also use the computer directly built-in "when identifying the other party's intention is to greet you when the corresponding solution" to eliminate the "independent generation of solutions" process. It is not even necessary to recognize the intent of the other side of the process, but to detect the other side of the sentence contains "some of the key words" automatically matched to "some kind of preset intention of the situation."

For example, I can set up a computer to put "grass mud horse" or other insulting vocabulary, the computer will be unified feedback "play Dan". Does the computer need to understand the meaning of "grass mud Horse" or "Play Dan" in this conversation? Is computer feedback inappropriate? Does the feedback of the computer belong to the intelligent behavior of "independent thinking" or the logical behavior of "according to the logical matching preset solution"? Are there any (meaningful) differences between the two methods?

Through these examples we can find that the human mind process is not particularly mysterious and difficult to explain the place, the human brain and computer rely on logical rules to process information. In fact, all the behavioral outcomes (the intelligent response) that humans exhibit can be logically explained. Humans "look more intelligent than animals" simply because the human brain's "performance" is much more powerful than that of the animal brain-the human brain is able to identify many complex things and respond to all the situations that come with "its own cognitive logic".

It is also true that people are often confronted with the "leap-thinking" of "A thought that pops up in the mind" or "relies on intuition to bypass logical thinking to get answers to questions", but this is an illusion:

The process of "experiencing inspiration" is when people do not realize that "certain thoughts that run silently on the back end" have come to a conclusion, it's like you don't have to consciously control your walk, but when your foot is tripped by a rock and you suddenly realize that "you're walking, you need to avoid being tripped and bumped."

"Encounter Intuition" is actually the brain "intentionally or unintentionally lazy" behavior, that is, according to some key information of the problem without "the process of dismantling critical information" directly in the "database" to match a result, and then put it directly into your mind, The results don't go through your brain's scrutiny like any other problem, which is why "intuition" sometimes (happens) to be right, but most of the time it's the root cause of the error.

To delve further, "inspiration or intuition" is a manifestation of the brain's "performance" limitations:

The brain uses this approach to reduce unnecessary energy depletion (similar to computer programs that can run in the background in a Windows system, but computers do not reduce performance losses, just to fit in with human behavior, allowing people to "focus on something that matters most to him." The human brain cannot keep its mind running in accordance with logical rules, because that would make "the brain or the people themselves" unbearable, so the brain also has its own "performance limit", people can not just think of "need to go through a lot of logical operation of the problem" to come to a "final conclusion", The brain tends to take a compromise approach to stop thinking about a problem and avoid letting the brain reach the limits of its performance.

The limitation of the computer means that it is limited by the "Hardware Performance", and the computer cannot "arrive at the final conclusion" for a "problem that requires a lot of logical operation" in "the stipulated time". Although the computer can follow the logic rules to a question "infinite thinking" down, but for some "really difficult problem", it is likely that there is no "meaningful to the human solution."

(and also to reduce performance loss) so we need to use "program algorithms" to limit the behavior of the computer (that is, to teach it in what circumstances it can compromise the knowledge), so that it's not "frozen because of the exhaustion of performance" or "wasting resources on a meaningless issue" (Is this a paradox?). We need to use the logic algorithm to let the computer solve the problem, we also need to use a logical algorithm to limit the computer on a problem indefinitely.

We can see "System error Please return" as a solution to this situation (computer performance limitations), but if the computer does not give the correct feedback on "Today is the day of the week", then why should it be used? Then it goes back to the original question: What kind of "intelligence" does a computer have to be called intelligent? or extend the problem to the outside: what is wisdom? Can intelligence be measured objectively? Is wisdom an objective existence or a subjective presence that must be manifested through biology?

The answers to these questions can actually be obtained from the objective fact of "the brain's performance limitations". As we speak of the brain's "compromise" on some "difficult" issues to reduce energy loss, the process is what we think of as intelligent behavior:

We say a chair, in fact it is wood, it is a combination of carbon atoms and other atoms, it is also the embodiment of some laws of physics. But for most ordinary people, just knowing that it is a chair is enough, if you delve into the question of why it is a chair, you will plunge yourself into a position of infinite thought and infinite computation, so we must avoid this by some means known as "wisdom".

Wisdom makes logic stop at a certain node to produce a "compromise conclusion" (i.e. knowledge), which is the manifestation of logic running to a certain stage. Wisdom is always relative, intelligence is a kind of "subjective existence", so it cannot be measured objectively. Wisdom is actually equivalent to the physical law of "making stone stone, making wood into wood".

With such a cognitive premise, we can explain most of the problems we face:

First of all, wisdom does not exist, in the world only physical rules or things called logic can be called objective existence;

Then our planet is naturally formed in accordance with the rules of physics, and all "activities" on earth can be explained by physical rules or logic, until the birth of life;

How does life occur? There is no evidence to prove that life was formed by virtue of "an unnatural force of the outside world" (instead "the simple form of life that prevails in the universe" has been somewhat substantiated), we can only think of the birth process as "the natural environment of the local sphere to a certain extent will naturally produce life "The natural result;

The most obvious difference between life and non-life is that "all organisms seem to have a purpose", which is to survive and reproduce. But can survival and reproduction really be called "active (subjective) purpose"? Are they fundamentally different from objective physical rules? The planet is based on physical rules "or condensed (alive) or dissipated (dead)", if this condensation of a single planet cannot be seen as an "active purpose", then billions of planets are grouped together to form galaxies, thus allowing for a more enduring "presence", a planetary explosion dissipates, What is the difference between the process of living and reproducing the other planets that contain the elements of the planet that are again condensed? Therefore, from the macroscopic point of view, the life of the expression of the "survival and reproduction process" is not divorced from the physical rules or the category called logic;

If there is no essential difference between life and non life, how can we explain the existence of human beings? From a macroscopic point of view, human beings are the result of natural evolution of the fittest in the Earth's biological world. From the microscopic point of view, human beings are composed of "atoms in certain conditions of the composition of biological molecules" and then "in some conditions combined into a complex organism." Human beings, though complex, are not unexplained. Humans always take for granted that they are beyond the logic of some kind of "impersonal" existence, but in fact humans with a stone or a worm or a dog, there is no essential difference between;

All these things have to follow all the rules of Physics (logic), life and death and the condensation and dissipation of the planet are the concrete manifestation of this rule. Although seemingly powerful, they are also vulnerable to a mess, simply unable to escape the limitations of physical laws (that is, between "life and death" repeated rotation);

For example, we have been able to build aircraft rockets to other planets to explore, but when we face the small Ebola or aids simple life forms of threats, still helpless. In the traditional cognitive concept of "incomprehensible" ah, so complex aircraft rockets can be created, such a simple biological virus is unable to cope with? But this is the manifestation of the physical rule--no matter how intelligent or powerful the organism is, which cannot transcend the existence of the physical rule, which is also embodied in the fact that "one cannot objectively measure a machine's intelligence". Because Wisdom is a (slightly complex) manifestation of the Physical rules (logic), you cannot measure the physical rules themselves by using a physical rule based approach.

In conclusion, if we can accept that wisdom is actually a concrete manifestation of the physical rules, we can then explain what we encounter as "artificial intelligence, a life and inanimate phenomenon entangled together cannot be divided" situation.

Well, this article has been basically a complete expression of my point of view, and it is obviously very difficult to get people to accept that they are not special machines and may have intelligent views. It might be more helpful to understand this article if I have the ability to use it as a "mathematical model", but I can only put forward theoretical points of view, and the relevant validation work requires other competent people to do it.

Finally, I would like to give a preview of what I am going to discuss in the next article: If Wisdom is an impersonal reality, then what is the meaning of life (is it not to produce wisdom)? If we have no difference in nature from a stone, does our "subjective existence" (i.e. life and consciousness) have "some acceptable value (meaning)"? My next article will challenge these issues, please look forward to-"civilized decoding: The bottleneck of human civilization and the Way Out analysis"!

(Responsible editor: Mengyishan)