Xueba's computing system

Chapter 45: Tutor's Complaints

"The school's funding is a little insufficient. I'll ask Li Youcai if it's possible to buy some equipment."

After hearing what Professor Zhang said, Lin Yuan felt a little relieved.

It seems that although the instructor also feels that his proposal of "letting algorithms produce algorithms" is not very practical, he is still willing to support my efforts.

As for why Lin Yuan insisted on making up stories, it was because he couldn't just say "the system taught me all this".

The computing system can actively iterate the AI model itself when training it, just like a supercomputer with a brain. But this ability is beyond the current level of human technological development. Otherwise, Lin Yuan would not be considered a hacker.

Lin Yuan's biggest problem was not the lack of computing power in the system, but how to find a perfect excuse to avoid suspicion when he used the power given by the system in reality.

You know, what science and engineering fields lack the most is idiots.

In a novel, you can write "a cloud of sadness floats over the horizon", and liberal arts students will not be too obsessed with it. But among the group of science and engineering students, there will always be people who ask a question - how do you prove that the cloud is sad.

What is your theoretical basis? Do you have any experimental data?

Lin Yuan thought about this problem for a long time, and finally got inspiration from the characteristics of the computing power system itself.

What is the system?
If the system itself is a supermodel AI, then the process of the system running AI training is AI training AI.

So, Lin Yuan came up with the idea - let the algorithm generate the algorithm.

This view is definitely not wrong, and it even makes sense.

For example, someone put forward the view that if I can calculate every quark in the universe, I can know the past and future of the entire universe.

How can you prove that this is wrong? Because you can't really count every quark in the universe. And this statement seems quite logical, and even has the wild imagination of science fiction and the beauty of philosophy.

So, let the algorithm produce the algorithm. How can you prove that this is not possible?

Even if I cannot prove it now, even if I fail in the future, my failure does not prove that my theory is wrong, but only proves that I cannot prove the theory I proposed.

However, Lin Yuan just kept these forced thoughts in his heart. He didn't dare to really say them out loud, for fear of being slapped to death by his mentor.

Perhaps it was Lin Yuan's honest and resilient personality that made Professor Zhang regard his nearly absurd view as a stubborn bull. And when a person reaches a certain age, he may find every bull cute.

Because that wild bull might be my former self.

"This Li Youcai always cries poor." Professor Zhang slammed his phone on the table, bringing Lin Yuan back to his senses.

Professor Zhang just had a phone call with Dean Li, and the other party obviously did not agree with the idea of licking enough computing power equipment.

"Teacher, if it's inconvenient, forget it. I'll find a way myself." Lin Yuan didn't want to cause other troubles because of this matter.

"It's not just for you. The institute was short of equipment to begin with, and it should have been purchased long ago. Li Youcai's son only cares about political achievements and the number of papers published. It's useless to spend so much money on publishing papers every year. There is no economic output from so many SCI papers a year. The sons don't feel sorry for selling their father's land, what a bunch of prodigal sons.

Is Musk's Starship based on SCI papers? This is a misleading statement, confusing cause and effect. . . "

Professor Zhang rants on and on, even to the point of being so angry that he barely coughs. "Teacher, calm down." Lin Yuan didn't expect his instructor to be so aggressive. He feels that if he doesn't interrupt him, he might rant about something even more explosive.

"I can ask seniors to lend me some mining cards. Then I can find a distributed AI training framework and work together."

The so-called distributed AI training is to split a large task into parts and assign them to different machines. However, just like the division of labor among people, the division of labor among machines will also have management expenses. Even if Lin Yuan can borrow a bunch of mining cards, it is not a simple matter to make them work together and run normally.

What's more, Lin Yuan probably can't borrow a bunch of mining cards.

"Mining card?" Professor Zhang's calm emotions were ignited again. "The money spent on each graduate student's thesis is more than the money spent on buying equipment. And then you kids are asked to work on topics that are either out of touch with reality or have been done to death by companies.

Let's take the field of image recognition for example. Companies have already done OCR character recognition and other things like that, so what's the point of researching them?

I spent all day studying how to make light bulbs, under the pretext of understanding the principles.

There are only a few Einsteins in a hundred years. Can people of Einstein's level be cultivated?

Not based on social needs, not based on practical applications. That is why the scientific research results of universities in the field of AI are no longer able to catch up with those of enterprises.

Could it be that the Double 11 recommendation algorithm came from some university laboratory?

If you say this to their faces, they will all be arrogant and say that this is just an engineering application.

Then you ask them what their research results are. They show you data. The data is very beautiful. But if you look at the influential papers in the field of AI in the past 20 years, which one was produced by them... "

Lin Yuan stood aside and listened to his mentor's complaints, not knowing how to respond.

After a while, Professor Zhang finally vented his anger.

“Alas, as I get older, I can’t help but be outspoken.

What I just said is not entirely correct, you can choose for yourself. After all, you still have to consider graduating successfully first. "

Lin Yuan naturally understood that if he had even one percent of the rebelliousness that his mentor had just said, he would be finished.

Lin Yuan also understood that the tutor was complaining about the current state of scientific research in China by complaining about the university. The tutor did not hold the kind of public opinion that is commonly seen on the Internet, such as: focusing on the domestic university education system and the college entrance examination system. Instead, he complained about the problem of disconnection between universities and enterprises.

Judging from the instructor's words, he obviously attaches more importance to engineering applications, which coincides with Lin Yuan's interests.

In terms of his interest in academics and applications, Lin Yuan never seemed to hesitate, and seemed to have gone straight to engineering applications from the beginning. This was probably because he wanted to take the postgraduate entrance examination because he couldn't stand the hardship of delivering takeout.

But after the instructor figured it out for a long time, Lin Yuan still saw no hope of solving the problem of computing power equipment.

What was the point of the instructor asking him to go back and avoid Gu Xuan? It couldn't be just to watch his old angry youth performance.

"I was forced by Li Youcai and MLG. I have to go back to my old job... I am already old and still want to do this kind of thing."

Lin Yuan's eardrums were a little uncomfortable with the teacher's eloquent words, but he was able to detect them, "Teacher, what do you mean by returning to your old job?"

(End of this chapter)