Chapter 124: Research Projects of the School of Computer Science and Engineering

Deciding to try to prove Qiu Chengtong's conjecture, Xu Yuan did not show up in any exchange activities for the next few days. Basically, after dinner, he went to Qiu Chengtong's office to study differential geometry, algebraic geometry, multi-complex functions and other branches of mathematics.

Qiu Chengtong also really hopes that Xu Yuan can make up for this regret and pass on all of his related research over the years.

Whenever Xu Yuan encounters complex and difficult problems, he will also explain them patiently.

He treated him as his own student.

The more he interacted with Xu Yuan, the more he could feel Xu Yuan's terrifying talent in mathematics.

Every time he only needed a little guidance and instruction, Xu Yuan would be able to understand the knowledge thoroughly and even apply it to other situations. Even he envied such a talent for mathematics.

But this also proves that his decision was not wrong.

As long as Xu Yuan persists in his research, he may be able to put a successful conclusion to the Calabi conjecture.

The weekend came in a blink of an eye.

The week-long exchange activity is coming to an end.

Sunday, April 2005, 4, Institute of Mathematical Sciences, Chinese University of Hong Kong.

In the office, Xu Yuan sat at a side table, with a pile of information and manuscripts filled with mathematical formulas in front of him. However, at this moment, his expression seemed solemn and his brows were slightly frowned.

After a long time, he came out of the deep learning state. Before the panel disappeared, he looked at the progress bar again and breathed a sigh of relief.

-

Task: Proof of Qiu Chengtong's conjecture
Subject: Mathematics

Progress: 30%
Result: Not completed
"It really is a question that has troubled Professor Qiu for decades. It is not easy to prove this conjecture."

In the past few days, he has been madly absorbing various relevant knowledge and mastering methods in many branches of mathematics, but he has not been able to prove the Qiu Chengtong conjecture that the first Chern class is positive.

Based on past experience in solving difficult mathematical problems, it is necessary to innovate in a variety of methods.

However, if Qiu Chengtong knew that he had advanced 30% of the proof of the conjecture in just a few days, he would definitely feel it was incredible, although this was mainly due to his research results.

After all, it had been thirty years since he solved the Calabi conjecture that the first Chern class is negative zero.

Just as Xu Yuan was thinking about this, Qiu Chengtong pushed the door open and came in from outside and greeted him.

"Did you gain anything today?"

Upon hearing this, Xu Yuan quickly stood up and turned his head away, answering with a somewhat embarrassed look on his face.

“There has been no breakthrough progress so far. I want to combine differential geometry, algebraic geometry, and multi-complex functions to prove the problem, but it is much more difficult than I thought.”

"Take it easy."

"It's hard to solve a math problem overnight. Sometimes it takes several years of research."

"Not urgent."

Qiu Chengtong waved his hand to comfort Xu Yuan.

Xu Yuan's performance these days has far exceeded his expectations, and he is very optimistic about this student.

Then he walked to his desk and sat down, and then he thought of something and said something.

"Are you going back tomorrow?"

"Yes, Professor Qiu, our Director Liao has already arranged the itinerary." Xu Yuan answered truthfully when asked.

Deep down, Qiu Chengtong certainly didn't want Xu Yuan to leave, but Xu Yuan was not a student of the Department of Mathematics at the Chinese University after all, so he couldn't force him to stay and not let him go.

So he sighed and instructed: "Take all these materials, manuscripts and books back with you. If you have any questions, you can email me at any time."

"Don't worry, Professor Qiu, I won't give up on proving this conjecture." Xu Yuan nodded and assured.

Then the two of them discussed mathematical methods, and time passed by unknowingly.

the next day.

The Chinese University of Hong Kong held a special farewell ceremony.

Amidst the farewells, Xu Yuan and others drove to the airport, preparing to return to Qinghua University.

At this moment, we can see that everyone in the car has a happy smile on their face, and it is obvious that they have gained a lot from this trip.

Although the first lecture by Qiu Chengtong was more difficult, the subsequent ones were all of normal level and I did learn a lot.

In particular, Zhang Jundong has determined the topic of his graduation thesis.

At the moment, he can be said to be full of confidence and can't wait to return to school and start writing.

"This exchange was really worthwhile. It made me feel enlightened and I immediately decided on my graduation thesis. Now I am very confident about my thesis."

"Then congratulations in advance on completing your graduation thesis." Liu Yang listened to what Zhang Jundong said beside him and did not forget to agree.

After a week of getting along, the relationship between the two became more familiar.

Because Xu Yuan did not participate in exchange activities every day, Liu Yang and Zhang Jundong would go to lectures together to communicate with each other.

Zhang Jundong felt happy when he heard this, and at the same time, he did not forget to look at Xu Yuan who was sitting in front of him.

"The one who gained the most this time should be God Xu. After all, it was Professor Qiu who taught the class in person every day. Even PhDs don't get such top treatment."

"It seems that Professor Qiu wants Brother Yuan to prove Qiu Chengtong's conjecture." Liu Yang replied as if he had thought of something.

Xu Yuan did not participate in the exchange activities in the next few days, so everyone was naturally curious and asked about the relevant situation. Therefore, Qiu Chengtong guessed that this matter was not a secret.

However, after Liu Yang said this, he and Zhang Jundong next to him fell into silence.

The Qiu Chengtong conjecture is naturally not something they can touch, but as the person who proposed it, Professor Qiu Chengtong has not proved it for decades. Just thinking about it, you can understand how difficult it is.

I couldn't help but worry for Xu Yuan, it was really not easy to prove it.

As for the conversation between Liu Yang and Zhang Jundong, Xu Yuan, who was sitting in the front row, naturally heard everything.

He was naturally aware of the difficulty of this problem, as it was, after all, a Calabi conjecture.

Previously, he solved the problem of infinite intervals between Carmichael numbers and the overall smoothness of the solutions to the Monge-Ampere equation. These only involved the mathematical branches of number theory and partial differential equations, which were not too complicated in some ways.

But Qiu Chengtong's conjecture is completely different.

Professor Qiu Chengtong transformed the Calabi conjecture that the first Chern class is positive into a stability problem in algebraic geometry.

This involves multiple branches of mathematics such as algebraic geometry, differential geometry, multi-complex functions and metric geometry.

Just combining these branches of mathematics is a considerable challenge.

Not to mention any more innovation.

Moreover, this conjecture also involves string theory in high-dimensional space, which is a problem deduced by theoretical physicists.

But now we have to rely on mathematical calculations, which is indeed not easy.

Otherwise, the so-called algebraic geometry stability problem could have been solved long ago based on Qiu Chengtong's proof of the Calabi conjecture that the first Chern class is negative zero.

Thanks to Professor Qiu Chengtong's research results over the years, the proof progress has been advanced to 30%.

And there was no growth afterwards.

Compared with the first two solved mathematical problems, the complexity has increased dramatically.

Xu Yuan understood all this and accepted it, but he did not let it knock his self-confidence to the point where he had the thought of giving up.

Anyway, he is still young and talented. At worst, he can slowly study and try to prove it.

He believed that it would be resolved someday.

Maybe one day inspiration will strike and this problem can be solved in a very short time.

……

The journey was very smooth, and Xu Yuan and others returned safely to Qinghua University before dark.

Xu Yuan and Liu Yang had just arrived at the dormitory and were planning to share what had happened in the past few days with Li Zhenya, Wang Haitao and others, when Xu Yuan suddenly received a call from Tang Shihong asking him to go to the office.

Thinking it might be related to this exchange activity, he didn't waste too much time.

After saying something to everyone, he hurried to the professor's office.

He soon arrived at the door with ease, raised his hand and knocked twice, and then he heard a familiar voice coming from inside.

"Please come in."

Xu Yuan pushed the door open and asked, "Teacher, is there something that asked me to come here?"

"No big deal."

"Sit down first."

When Tang Shihong saw Xu Yuan coming, he quickly asked him to sit down.

Then he asked about the results of this exchange.

"After studying with Professor Qiu for a few days, what have you learned?"

"Achieve a lot."

Xu Yuan responded truthfully: "Especially, he has mastered many methods in many branches of mathematics."

"I hope the trip was worthwhile." Tang Shihong nodded.

As soon as he finished speaking, he raised a new question: "How confident are you in solving Qiu Chengtong's conjecture?"

Xu Yuan was not surprised that Tang Shihong also knew about this.

He paused for a moment to organize his words, then slowly began to talk about Qiu Chengtong's conjecture.

"The stability problem of algebraic geometry involves many branches of mathematics, and it is not easy to solve it completely. However, Professor Qiu explained his research experience and some ideas to me, so I wanted to give it a try."

"In that case, it's up to you to decide, but don't put too much pressure on yourself."

After hearing this, Tang Shihong gave some advice but did not raise any objections like he did in his previous graduation thesis.

Ever since Xu Yuan showed his talent in proving the smoothness of the solution to the Monge-Ampere equation in his graduation thesis, Tang Shihong has had an inexplicable trust in Xu Yuan.

Even though he knew the difficulty of Qiu Chengtong's conjecture, he was still willing to believe that Xu Yuan could complete the proof.

The next second Tang Shihong thought of something and said, "Also, Mr. Yang is also optimistic about your solution to Qiu Chengtong's conjecture."

"Mr. Yang also knows about it?" Xu Yuan was slightly surprised.

He didn't expect that he had just returned from Hong Kong, and it felt like everyone in the school knew that he was going to challenge Qiu Chengtong's conjecture.

When he proved the related problems of Carmichael number and Monge-Ampere equation, no one knew about it. It was only after the paper was published that he attracted the attention of the academic community and was awarded the title of Prince of Mathematics.

Now that so many people know that he is studying the Qiu Chengtong conjecture, it would be embarrassing if he fails to prove it.

Tang Shihong did not keep Xu Yuan in suspense and immediately told him all the details.

After hearing this, Xu Yuan was at a loss whether to laugh or cry, and didn't know what to say.

It turns out that Professor Yang has always intended to invite Qiu Chengtong to join Qinghua and become a permanent member of the Qinghua Institute for Advanced Studies, responsible for related work of the School of Mathematical Sciences.

However, this matter has not been accomplished yet.

This time he tried to prove Qiu Chengtong's conjecture. If he succeeded, he could convince Qiu Chengtong to join Qinghua.

Therefore, Mr. Yang is also very optimistic about him.

After figuring out the whole story, Xu Yuan could feel a lot more pressure on him.

At the moment, it seems that everything must be proven.

"Don't worry, mentor. I will do my best to study it."

Finally, Xu Yuan stood up and made his attitude clear. After saying this, he did not linger any longer. He opened the office door and left.

Tang Shihong looked at Xu Yuan until his back completely disappeared from his sight.

He shook his head and sighed for a long time.

"I don't know whether it is a good thing or a bad thing for him to have this kind of confidence."

What satisfied him most was the confidence that Xu Yuan always possessed, and he even envied him very much.

But with this kind of confidence, if the Qiu Chengtong conjecture cannot be proven, it may have the opposite effect.

The whole person's self-confidence was hit.

You should know that Xu Yuan has never suffered any setbacks since he started learning mathematics, thanks to his mathematical talent.

However, the situation cannot be changed now, so we can only hope that Yu Xuyuan can complete the proof.

-

Time flies by like a flash.

Soon it was the end of June.

It has been two months since Xu Yuan returned from Hong Kong, and his schedule has been very full during this period.

In addition to continuing to learn new knowledge and attending classes, he spent the rest of his time in the discussion room studying the Qiu Chengtong conjecture. During this period, he also sent emails to Professor Qiu Chengtong to ask for advice on any questions he had.

Unfortunately, the progress bar has not shown any growth in two months.

Still remains at 30 percent.

Although he tried many innovative methods, the progress bar would drop every time he substituted the calculations.

This means that his methods are wrong or cannot be used to prove Qiu Chengtong's conjecture.

This actually saved him a lot of time.

You can continue to try and error.

Because under normal circumstances, if you want to determine whether the proposed method is effective for the result, you need to substitute it into the calculation and verify it.

I don't know how much time was wasted during this whole process.

Fortunately, Chen Yuran would come over from time to time to study with him and encourage him at the right time, so even though his progress did not increase, his fighting spirit became more and more excited.

I secretly struggled with this conjecture and was determined to solve it completely no matter what.

This puts an end to the Calabi conjecture thirty years later.

That morning in the discussion room of the School of Mathematical Sciences, Xu Yuan picked up the draft paper on the table and frowned slightly.

"As expected, it still doesn't work?"

He muttered to himself, then crumpled up his fist and threw it into the trash can.

He tried to solve it using partial differential equations, but the progress bar on the panel still dropped significantly.

This shows that this method is wrong.

By now he was certain that he could not solve the Calabi conjecture that the first Chern class was positive by using partial differential equations, and he still had to find ideas from differential geometry and multi-complex functions.

Originally, he thought that since Professor Qiu Chengtong could solve partial differential equations through the concept of curvature in geometry, thereby proving the two cases of Calabi conjecture when the first Chern class is negative or zero, then perhaps this method could also be used to solve the Calabi conjecture when the first Chern class is positive.

It is nothing more than making appropriate innovations or optimizing on the original basis.

However, he got wrong results after many attempts, which shows that this method does not work when the first Chern class is positive.

No wonder Professor Qiu Chengtong only solved the two cases when the first Chern class was negative or zero.

Perhaps because he had been studying this conjecture every day for two consecutive months, Xu Yuan could clearly feel that his thinking was beginning to become rigid.

It's hard to come up with clever ideas through collision.

I thought I should find something else to study to change my mind, maybe I can get twice the result with half the effort.

He remembered that he had previously proved a problem related to the Monge-Ampere equation and was invited by Professor Qi Tongxiang to serve as a judge and give lectures at USTC. This gave him a new idea that significantly increased the progress bar.

When Xu Yuan thought about this, he couldn't help but mutter to himself while pondering.

"Do we have to hold an open class to talk about Qiu Chengtong's conjecture? I'm afraid not many people will be interested."

Just as he was hesitating whether to apply to the hospital, the sudden ringing of the cell phone in his pocket interrupted his thoughts.

I casually took out my cell phone and looked at the small screen, and immediately saw Professor Tang's name appear on it.

Although he didn't know what Professor Tang wanted to say, he didn't dare to delay and quickly pressed the answer button.

As soon as I put the phone to my ear, I heard a familiar voice coming from the receiver.

"Xu Yuan, are you free now?"

"I'm fine." Xu Yuan answered without hesitation.

Tang Shihong: "Then come over here for a moment. I have something to tell you."

"Good mentor."

Xu Yuan naturally didn't think too much about it. He answered, hung up the phone, and set off for the comprehensive experimental building where Tang Shihong's office was located.

……

As soon as he entered Professor Tang's office, he found that not only Professor Tang was there, but also Professor Yao Qizhi from the School of Computer Science was sitting on the sofa next to him.

A thought suddenly popped into my mind.

"Could it be that what Professor Tang said on the phone is related to the School of Computer Science?"

Although he had never attended Professor Yao Qizhi's lectures, he was not unfamiliar with this big man.

Last year, at the invitation of Mr. Yang, he gave up his overseas job and officially joined Qinghua Institute for Advanced Studies as a full-time professor, responsible for teaching computer majors and related project research.

Moreover, he is the only Chinese scholar to have won the Turing Award and enjoys high authority internationally.

"What are you still standing at the door for? Come in and sit down." Tang Shihong looked at Xu Yuan standing at the door, stopped what he was doing, and asked with a smile, "How is the research on Qiu Chengtong's conjecture going recently?"

"I have no clue yet..."

Xu Yuan answered Tang Shihong's question with a smile, greeted Professor Yao Qizhi politely, and then sat down on the sofa obediently.

Yao Qizhi asked with a smile: "It takes courage to prove such a difficult mathematical problem."

Xu Yuan said modestly: "I just want to give it a try." He was absolutely right. No matter how difficult the mathematical conjecture is, you should always try to prove it.

Otherwise, how would you know that you are not capable?

Even if they fail, it's no big deal for mathematicians.

"Teacher, is there something important that made you suddenly call me here?" Xu Yuan changed the subject in time.

Hearing this, Tang Shihong did not directly state the matter, but instead raised a question first.

"What are your plans for this summer vacation?"

"I plan to stay in school during the summer vacation." Xu Yuan thought for a while and said, "Qiu Chengtong's conjecture still needs time to study."

After the summer vacation, he will be an official doctoral student in science. In addition to the Qiu Chengtong conjecture problem he is currently working on, he also has other academic matters that need to be dealt with.

Compared to going home to study, it is undoubtedly more convenient to stay in school.

He has also called his family about this matter, and his parents expressed their strong support.

In addition, her elder sister Xu Ying will be a senior when school starts and will also face the problem of internship and job finding.

However, he contacted Zhang Yuhui of Luyang No. 1 Middle School in the provincial capital about this matter, and planned to let his sister become a teacher in the provincial high school in the future.

At least it is convenient and stable.

While Xu Yuan was thinking about this, Tang Shihong and Yao Qizhi next to him both smiled, and seemed very satisfied with the answer.

Then he continued to talk about the specific situation.

"It's not me who asked you out this time." Tang Shihong shook his head and said, "Professor Yao is studying machine learning recently. The Monge-Ampere equation of optimal transmission theory that you proved can be used in algorithms, so Professor Yao wants to invite you to join this project during the summer vacation, responsible for solving mathematical content such as algorithm functions and probability Gaussian distribution."

After listening to Tang Shihong's words, Xu Yuan finally understood what was going on.

First of all, he must be familiar with machine learning algorithms. In recent years, the number of research teams in China has gradually increased.

As a core component of artificial intelligence, machine learning can play a role in many fields such as image processing engineering applications.

In addition to his remarkable achievements in mathematics, the optimal theory of the Monge-Ampere equation he studied is an important part of machine learning, so it was reasonable for Yao Qizhi to ask him for help.

"Optimal transport theory is a bridge between geometry and probability. It provides powerful tools for modeling probability distributions and measuring the distance between probability distributions using geometric methods."

"It can be said that optimal transfer theory provides a new perspective for explaining the principles of machine learning."

"It provides new guidance for improving machine learning algorithms."

"When I heard Professor Tang recommend you to come over to help, I was so happy that I wanted to take you away right away."

Yao Qizhi did not remain silent at this time. He spoke with great excitement, and his words showed his praise for Xu Yuan.

Hearing this, Tang Shihong smiled even more brightly: "Xu Yuan is indeed the most suitable candidate for our School of Mathematics."

Although Xu Yuan was happy to be praised by Professor Yao, he still prepared to be modest.

But before he could open his mouth, Tang Shihong's voice suddenly came to his ears again.

"I know that you have been focusing on the research of Qiu Chengtong's conjecture recently, so it is up to you to decide whether to go or not."

"This project won't last too long. Several doctoral students will be responsible for the algorithm program. You only need to provide support in the underlying logic and mathematics." Yao Qizhi was worried that Xu Yuan would refuse, so he hurriedly added some explanation.

Xu Yuan does not find it difficult to research and improve machine learning algorithms. It is a relatively easy project for him.

Besides, you only need to be responsible for mathematics.

It just so happened that he had no ideas for his research on Qiu Chengtong's conjecture recently, so it would be a good idea for him to go to the School of Computer Science to help change his mind.

Besides, the optimal transmission theory connects geometry and probability, which can also help him study geometric problems. Maybe he can get some inspiration and find a way to solve Qiu Chengtong's conjecture.

The thought stayed here, and without any hesitation, Xu Yuan immediately informed Professor Yao Qizhi of his decision.

"I would like to go."

When Yao Qizhi heard this answer, his face was immediately filled with smiles.

"Great."

"It's better to go today than to wait for a better day. I'll go today and get familiar with the place."

Having said that, he stood up from the sofa and spoke to Tang Shihong in a joking tone.

"I won't bother you anymore since you have been taking so long today. As for your precious student, I will borrow her."

"Professor Yao, you don't have to be so polite." Tang Shihong said with a smile.

……

Xu Yuan had nothing to do, so he simply followed Yao Qizhi to the project site of the Institute of Planning.

This was Xu Yuan's first time back to the School of Computer Science. Looking around, he found that the environment was no different from that of the Department of Mathematics, except that he met slightly more girls on the way.

Although it can't compare with the Journalism Department, it can definitely beat the School of Mathematics and Science.

After all, computer science has been very popular in recent years, and the number of applicants is relatively large.

“Have you been here before?”

At this time, Yao Qizhi randomly picked a topic to ask, hoping to make the atmosphere less boring and awkward.

Xu Yuan answered truthfully: "To be honest with you, Professor Yao, this is my first time coming to the School of Computer Science."

"Since you joined this project, you have come here more often. To be honest, I have not been in Qinghua as long as you have." Yao Qizhi continued talking to Xu Yuan like an elder, with some humor in his words.

And what he said was indeed correct. After all, the other party was invited to join Qinghua as a full-time professor last year.

Calculating the time, it is naturally not as long as the time Xu Yuan stayed in Qinghua.

Of course, this was just a small episode, and soon the two of them arrived at the Institute of Network Technology of the School of Planning and Design.

"We're here."

Following Yao Qizhi's reminder, Xu Yuan followed him inside and found that the interior was no different from the teaching building.

It's just that the room is filled with computers.

After Yao Qizhi brought Xu Yuan to one of the rooms, he found a tall and thin doctoral student wearing a pair of black-framed glasses.

"Xiao Guo."

"This is Xu Yuan from the Department of Mathematics. He will be responsible for the mathematical content and will help you improve the algorithm. First, take him to get familiar with the environment and other members."

"Okay, Professor Yao." Guo Zhihang replied with a smile.

Yao Qizhi entrusted Xu Yuan to Guo Zhihang and left here temporarily.

Guo Zhihang habitually pushed his glasses down on his nose, then took the initiative to extend his hand and introduced enthusiastically.

"My name is Guo Zhihang. If you don't mind, you can just call me Brother Guo."

"Xu Yuan."

Xu Yuan smiled and shook his hand: "Brother Guo."

"Don't be polite when you're here. We've heard of your name for a long time."

"You are a celebrity in our Qinghua."

"I didn't expect Professor Yao to call you over. Our machine learning algorithm will definitely be improved this time."

Guo Zhihang was very enthusiastic. Even though it was the first time he met Xu Yuan, he didn't feel any strangeness at all. If someone didn't know him, they might have thought they had been classmates for many years.

After Xu Yuan withdrew his hand, he was a little surprised. He didn't expect that his name had already spread in the School of Finance.

But after all, the other party was a doctoral student who was older than him, so Xu Yuan chose to be a little humble, treating himself as a newcomer and not showing off his math grades.

"Brother Guo, thank you for your praise. As long as I can help you, that's fine."

"Let me introduce you to a few other people." Guo Zhihang said with a brighter smile on his face when he heard this.

As soon as he finished speaking, he dragged Xu Yuan towards a few people who had stood up not far away.

"This is Zhu Minfeng. You can just call him Senior Brother Zhu. He is also a doctoral student like me."

"Senior Brother Zhu." Xu Yuan stretched out his hand and shook hands with the other party.

Zhu Minfeng was obviously not unfamiliar with Xu Yuan, as could be seen from the excitement on his face.

"Brother Xu, I just read your paper on the smoothness of the solution to the optimal transmission Monge-Ampere equation not long ago. Both the proof method and the process are exquisite."

"I originally thought you didn't have time, but I didn't expect Professor Yao to actually ask you to come over."

After hearing what Zhu Minfeng said, Xu Yuan couldn't help but feel curious.

Then he casually asked, "Brother Zhu, are you also studying partial differential equations?"

"I studied mathematics for my undergraduate and master's degrees." Zhu Minfeng answered with a smile when asked.

Xu Yuan suddenly realized: "So that's how it is."

In addition to the two doctoral students Guo Zhihang and Zhu Minfeng, there are two other graduate students in the team.

One male and one female.

After introducing each other, they also called Xu Yuansheng "Senior Brother Xu".

Afterwards, Guo Zhihang gave a detailed introduction to their project and current research progress.

Under the guidance of Professor Yao Qizhi, several of them used the generative adversarial network method to initially complete the machine learning algorithm, but the results of the algorithm performance evaluation were very poor.

It is far from achieving the expected goals.

Their original idea was to develop a machine learning algorithm with better performance than the current level, apply for national funding support in the form of a paper after the project is completed, and then conduct more in-depth research. From the practical application of machine learning to the research of artificial intelligence, it can be said that they have a long-term goal.

However, based solely on their current algorithm performance, they will certainly not be able to get support from national fund projects.

So basically, we placed our hopes on Xu Yuan, hoping to improve the algorithm performance by optimizing the mathematical function formulas.

Xu Yuan had only been involved in hash function cryptographic algorithms before, but he felt a sense of freshness about machine learning algorithms.

As soon as I was assigned a seat in front of a high-configuration computer, I couldn't help but rub my hands and couldn't wait to start my work.

"Junior Brother Xu."

Guo Zhihang came over again at this time, holding a USB flash drive in his hand.

“This is information that feeds into our machine learning algorithms and helps you refine your math.”

"Okay, I'll study it first." Xu Yuan reached out, took the USB drive, and nodded.

After doing this, Guo Zhihang did not leave, but instead raised another question.

"By the way, Junior Brother Xu, do you know anything about computer languages ​​and modeling?"

When Xu Yuan heard the other party asking this, a mysterious expression suddenly appeared on his face. The corners of his mouth slightly raised and he uttered two words with a smile.

"Understood."

Not to mention that there was a computer in his discussion room and he often studied it, his computer skills in his previous life were also very advanced.

It is no exaggeration to say that there is no problem even if he is responsible for the entire machine learning algorithm alone, it will just take a little more time to complete it.

"Junior Brother Xu is only responsible for the math content. Just know a little bit about computers." Guo Zhihang took over the conversation and responded: "As for the math function formula and modeling, don't be too anxious. Junior Brother Xu, just take your time. If you have any questions, feel free to communicate with us."

"I understand, Senior Brother Guo." Xu Yuan nodded when he heard this.

But although he said this, Xu Yuan knew that Guo Zhihang and Zhu Minfeng must be very anxious.

After all, it is difficult for them to improve the algorithm before new logical formulas and modeling are developed.

In other words, they have to wait until Xu Yuan completes his task before they can carry out subsequent work. Mainly because since they have invited Xu Yuan to participate in the project, they certainly cannot continue to use the previous modeling.

Besides, the previous model evaluated such a low performance, which in a sense was a failure.

Rather than optimizing it, it is better to start over.

And in fact, Xu Yuan did just that. After reading the information in the U disk, his brows were furrowed. He didn't expect that there were so many problems in it.

At least he can give a better choice with just a glance at the algorithm formula model used by Guo Zhihang and others.

No wonder the algorithm's estimated performance is so poor.

After research and consideration, Xu Yuan decided to use a combined neural network as the algorithm framework.

The principle is that each neural network module can be regarded as a brain of the machine, so this algorithm mainly makes each brain smarter through confrontation between brains.

Intelligence means that each neural network module is adjusted to enable the entire system to complete the designated task.

For example, image processing and speech recognition.

For this automation model, the most critical things are input and feedback as well as overall performance.

Because the input mechanism is randomly distributed, unlike the previous input which is labeled data, this means that this algorithm is a product of unsupervised learning, which is the most difficult and practical area to enter into deep learning.

Xu Yuan believes that once this machine learning algorithm is completed, it will put China in a leading position in machine learning research.

However, since the model involves a loss function and needs to use a mathematical optimal transmission framework instead of the neural network, it takes a certain amount of time to complete the calculation.

……

that's it.

After Xu Yuan officially joined the machine learning project of the School of Design, he would go to the Institute of Design to construct models during the day.

At night, I would go back to the discussion room to look for new proof methods.

Their living conditions can be described as regular.

In the blink of an eye, it was July. With the arrival of summer vacation and the school closing, the number of people on campus suddenly decreased a lot.

However, this did not have any impact on Xu Yuan. After more than half a month of hard work, he finally completed the model.

The overall expression is still the optimal solution.

However, when Guo Zhihang, Zhu Minfeng and others heard the news, their eyes widened instantly. They never expected that Xu Yuan had completed the algorithm model so quickly.

Tuesday, July 2005, 7, Institute of Network Technology, School of Computer Science and Engineering.

"Junior Brother Xu, you are not kidding us. Have you really finished all the algorithm formulas and models?"

In the project office, Guo Zhihang took the USB flash drive handed to him by Xu Yuan, his face still full of disbelief, and he couldn't help asking again and again.

The main reason is that this is completely different from what they knew. I remember that it took them several months to do this work at the beginning, unlike Xu Yuan who finished it in just half a month.

Zhu Minfeng, who was standing next to him, was worried that the model produced in such a rush would not perform well, so he added a sentence after a slight hesitation.

"Or Junior Brother Xu will optimize it. We are not in a hurry."

The two graduate students standing behind nodded vigorously although they did not speak at this time.

After listening to everything Guo Zhihang and Zhu Minfeng said, Xu Yuan didn't know whether to laugh or cry, thinking that his efficiency being too fast had actually become a problem.

But this cannot be blamed on him. After all, efficiency will be greatly improved after entering the deep learning state. In addition, he has no idea how to prove Qiu Chengtong's conjecture at night, so he will take out the project of the School of Computer Science to complete.

So the summer vacation had just begun and he had already completed his task.

It’s not that he is overconfident, but he is still quite confident about the work he has completed.

There is no need to optimize anything anymore.

At least in the current research on machine learning, its algorithm performance can definitely be regarded as leading.

So he had to confirm again: "Don't worry, brothers, we can wait until the algorithm comes out to evaluate the performance to see if it can meet the requirements."

When Guo Zhihang and Zhu Minfeng heard what Xu Yuan said, they hesitated for a moment but finally agreed.

After all, they invited Xu Yuan to help, so there was no way to question the results now that they hadn't verified them. Even if the final evaluated performance still didn't meet the requirements, it would only be a waste of time and could be accepted.

"OK then."

Zhu Minfeng said this, then turned to the other three and uttered two words in a deep voice.

"start."

After that, they all returned to their computers and began to complete the algorithm program based on Xu Yuan's model.

Originally, Xu Yuan wanted to help with programming, but was righteously rejected by Zhu Minfeng and Guo Zhihang.

Obviously, I don't believe in his computer expertise.

Xu Yuan was not angry about this, and continued to study the problem of solving the stability of algebraic geometry.

During the past half month, while helping with the School of Design's project, he had some ideas in mind, but he could not really grasp the key points. This situation made him feel a little annoyed.

Fortunately, Xu Yuan had no mental problems. He just studied and improved himself while researching mathematical branches and methods, and did not force himself to prove Qiu Chengtong's conjecture.

As time went by, with the efforts of Guo Zhihang, Zhu Minfeng and two other graduate students, Xu Yuan's machine learning algorithm that replaced the neural network with the optimal transmission framework was finally completed.

At this point, all that remains is to evaluate the algorithm's performance to determine its overall level.

On the day of the algorithm performance evaluation, in addition to Xu Yuan who was present at the scene, Guo Zhihang and his team also specially called Yao Qizhi to come over for the evaluation together.

Although everyone is concerned about Xu Yuan's amazing efficiency, they naturally have high hopes that the evaluated performance can meet expectations after investing so much effort and energy.

……

"Xu Yuan, have you ever studied machine learning before?"

That morning, when Guo Zhihang and his team were evaluating the algorithm, Yao Qizhi focused his attention entirely on Xu Yuan, not hiding the joy on his face at all.

Naturally, he looked at the formula model completed by Xu Yuan and was attracted by the advanced concepts therein. He did not expect that a mathematical optimal transmission framework could be used to replace the neural network.

Because it is different from the current mainstream machine learning algorithms, it is normal for Guo Zhihang and others to have doubts.

But he feels that the performance values ​​evaluated by this algorithm may bring a huge surprise to everyone.

Xu Yuan didn't think too much about the question and shook his head and replied: "I have only studied hash function algorithms."

……

(End of this chapter)