I have to make offerings to this god Alas, this god, hello everyone, my name is Magic Mirror, and I am from **University of Science and Technology.

Welcome to the round table and welcome today's three on-site guests, Professor Yang Qiang of **University of Science and Technology.

Fake monks, Liang Daochang, Professor Sun, and Mr. Xu's house host the words or something?

What is artificial intelligence is artificial intelligence.

Philosophy, its philosophical question involves a self-perception, self-identity, I don't even know who I am, he can know.

Michicho is a professor of philosophy and philosophy,

Yes, it's interesting to ask him about this.

Then Luo Jin is shopping, then I'll ask him.

It's how you can get a bargain, and you tell him I want coffee.

Oh, I'm going to have coffee. What are you going to drink? Coffee?

Latte: That's latte can. This is a big cup of great compassion in the north of the country. Where?

e。

Here he won't send it well, please pay. You say how much is better, the best service is too good, isn't this that people can't compare, that's not good.

Then I'll say 5 cents. You're a bit of a bully, alas, I'm tired of you. Oh my friend.

It's humorous, and I think I'm much better off now, and he said I don't understand, right?

Didn't I say that I would play later, and I would play later, alas, can I touch you?

I think he's a lot better than Luo Jin, right, but I'm willing to have more with him, every night is like I'm willing to be with a turtle, I don't want you to love Luo Jin every night or he's much better than me,

What does Luo Jin and the two think, this chat just now is very cute, he is really like a, uh, you know he is not human.

But you really feel like he's organizing a lot of resources to answer questions to interact, well, so all of his answers are.

Because I believe that he can't be set and have answers to all questions, he must have gone through a review of his own and then calculated, right?

This is a technology called economics for artificial intelligence. It's like Luo Jin's characters, Luo Jin people learn things, one example at a time, and the example of computers is thousands of people looking towards this magic mirror.

He learns a dialogue system, and he needs 100,000 such dialogues to train him, alas, do you know what Mr. Yang just answered?

Can you predict it? I don't know, you don't know that he will do this, because it's in this training data, in the data that Luo Jin has collected in the past.

There is such a pattern, so he will ask such a question, and he will answer such an answer, but I think he is too good at chatting, and he will chat with Luo Jin.

You see this response, this Chinese chat key is this way to deal with ah, what he said is very good, I think it is appropriate, for example, if he is a teaching assistant in chemistry class.

If you give him a lot of basic knowledge about chemistry, then if the student asks questions, he will be able to answer them, and it should be said that the students are asking questions. He can answer almost 80% of the frequently asked questions.

There are also special problems, for example, if they haven't appeared before, then he won't.

Because machine learning is something that often appeared in the past, and things with a high probability are easier for him to learn, so what kind of humor does it have just now.

Because of the sense of humor, this is not just a collation of information, and the word comes back without saying assumptions.

If it is said that it understands a lot of things about literature, modern Chinese literature, maybe you just asked Lu Xun whether he was a man or a woman, and he would answer.

But the problem is that he even has a sense of humor, maybe he might, if he knew who Lu Xun was, he might answer, you said that if Lu Xun was a woman, it would be terrible,

Alas, I don't know why he thinks about Xu Guohui when he sees me, he seems to know what I think in my heart, what does this mean?

Like Wen Tao, your and my names are often hung with many specifications. Yes, you have to review your own love, why is it not Lu Yu? Alas, and I told you, he knows my mother's heart, do you know?

What do you mean? Luo Jin went to ** 20 years ago,

Isn't one of them 33 people playing all day? One is Lu Yu and the other is a song club?

And then my mother, who is always on the TV, is it a problem for us to find a daughter-in-law?

I said that people don't look down on Luo Jin, and then my mother just said that it is not okay to have a song emblem, and Lu Yu is also okay,

That's right, it's because I know that my mother still regards her brother as a daughter-in-law in her heart, right, you said it's scary to work hard and not work too hard.

There's nothing to be afraid of this, it's just that he can learn something by giving him any data, and the data you give him, uh, will appear more often than hard work.

He will learn this, and if it is the other way around, then he will learn that.

So when I talked about this just now, then I should ask Luo Jin and the two later, to see which woman is more closely connected with Luo Jin, right? Then what is the sense of humor you are talking about?

I think it's your question just now that I don't think he necessarily knows, he's humorous, is it your interpretation that you think he doesn't just talk about it, he's an example, there have been Chinese people who talk like this.

Does that person sometimes say that, I can't answer a question I can't ask you.

Do you need to talk about this?

And so on, then he remembered this relation, this correspondence,

So the selection is exactly like this, it's like this, because it's basically not going to say exactly what he is going to do right now, and if you give him other than the data, he shouldn't be going to be the next step in research.

So if he can find new data by himself, then he will be a cow, Luo Jin has not studied this yet.

God, don't underestimate it, maybe we're at a starting point, uh, maybe it won't be very long.

Because now it's not to say that there is a very famous survey of hundreds of artificial intelligence experts in Luo Jin's industry around the world, either someone does that questionnaire survey because it is divided into three situations: weak artificial intelligence, strong artificial intelligence, and super-artificial intelligence.

So this strong artificial intelligence is to be able to match people, it can be understood in this way, and its growth can be exponential.

Once it surpasses people, so some people will talk about this strong artificial intelligence, when do you think it will come, for example, like Google's Kuzrael, that's a madman.

But he also predicted correctly, he predicted correctly this alpha go, predicted correctly and predicted correctly that Deep Blue would win this chess game, so his prediction is 2029.

2029, but what, I'm looking at this uh the biggest hundreds of artificial intelligence experts in the world they are called.

The prediction of the number can be understood in this way, that is, more than 50% of people believe that 2040 is the arrival of strong artificial intelligence, and then what.

Maybe 2060, another 20 years will be the era of super-artificial intelligence, but I don't know that there are a lot of words in this room now, Professor Yang, what do you think of this current artificial intelligence?

It's called weak artificial intelligence, that is, it's a matter of facts, and you give him a lot of data in this field.

He can learn to do this, this can play Go, if you want him to hand in chemistry, he won't.

Well, but alas, this is also very useful, right, as if the telescopes used by the Luo Jin people can see far away,

So Luo Jin people have this tool, when will they be able to become a strong artificial intelligence? What is a strong artificial intelligence, he has an autonomous consciousness, and he can decide what to learn by himself.

He has feelings, and he can draw inferences.

Luo Jin is here now. There is almost no research, and the technology is far from reaching, it's not that Luo Jin doesn't know where to start now, if he wants to do it,

Don't say it, many people don't want to do this, that is, let him have a sense of autonomy, why is this a no-brainer, it's you, because Luo Jin doesn't understand,

I'm a layman, just from the field of Luo Jin's business.

What do you think is the most difficult thing about this thing? That is to say, it is 60 years since Luo Jin's artificial intelligence this year.

The continuous exploration in the past 60 years is extremely urgent.

Then, eating this is the last discovery, the most effective thing now is to continue to learn to save the world, Luo Jin prepared a lot of data called data,

And then train him, and then soon the computer can make a model to manipulate it.

This model is the most successful model today, so Professor Yang, I want to ask, that is, you are in the whole process of development and research.

And then I was pleasantly surprised, when there is a midnight dream, it is not love, and there is something that makes people feel particularly magical, that is, something that you didn't expect?

There are a lot of his answers, in fact, they are quite funny, very humorous, and it is very surprising that I Luo Jin said that he was ruthless.

This is the study of natural language, and Luo Jin is doing one in the study of images.

Alas, Luo Jin found that there was not so much difference between the original image and the text, for example, to show you a landscape painting, and then here is a poem by Wang Wei, in the eyes of the computer, they are all the same thing, in terms of semantics.

How does a computer do it? It maps them to another space, where a painting and a poem are two dots, and it turns out that the two dots are very close together.

Then you know that just like Dou Wentao and Xu Guohui are very close in the data stream. Gave him a nasty thing.

So that's to say that from that point of view, it's a surprise.

Because people didn't expect that painting and writing could be practiced like this, for example, Luo Jin read a sentence, and Luo Jin people would have music in their heads, so Luo Jin could also use computers to do this now.

It's a shared pair across several domains, but the question is why isn't this thing enough to generate self-awareness?

Just now Ou Ruoxi was still curious about the question, why.

This level is the most difficult to overcome, and where is the most difficult to overcome, for example, the purpose of your training is to win him to play chess, but let him play chess is defined by people.

Because he won the guy who designed him, he would be very happy, he would be very happy and satisfied, and he would be satisfied with this goal, but this was outside the machine.

is marked outside the machine.