Huaxia, Northern Xinjiang University.
In the conference room on the third floor of the NLP R&D Center, which has the same signs as the Confidential and State Key Laboratory in Xilin University of Technology, a middle-aged man in his fifties was explaining a ppt on the stage.
“…Thanks to the cooperation with Youwei, we have integrated the processing of the crowd intelligence NLP part into our intelligent training model. At present, we have effectively solved the problem of our intelligent model in the field of large-scale collaborative operations of UAVs. Ability to understand unique military terminology and professional language.
But now we still encounter many problems. For example, under special circumstances, the judgment of encrypted semantics, and the text noise caused by some non-standard terms in emergency situations still pose a great test to the stability of the entire system.
Secondly, the response speed is still not ideal in processing multi-modal data such as images, audio, and video, especially the denoising process. This also directly leads to insufficient cooperative combat capabilities of large-scale drone groups and untimely intervention in the system.
According to our sandbox simulation, when a drone group exceeds a certain number, it can easily cause large-scale non-combat losses. At the same time, long-term data analysis still fails to meet expectations. The analysis of information and trend changes on the battlefield over a certain time range is not ideal.
Therefore, our next research and development work will focus on the following aspects. First……"
…
As the name suggests, the Beijiang NLP R&D Center is engaged in military natural language processing.
The most important task of the current R&D center is to provide corresponding supporting equipment for future drone clusters.
In a sense, it was similar to what Su Mucheng wanted Qiao Ze to do, but the requirements were higher and the fault tolerance rate was lower. It also has to adapt to an environment that is far more complex than civilian use, and the degree of difficulty is naturally higher.
Fortunately, this is mainly future-oriented technology.
At least at this stage, the requirements for research and development time are not that urgent yet, and the current technical indicators are sufficient for the time being.
But then again, large-scale drone swarm combat technology requires more than just the development of nlp. Many research institutes across the country are doing relevant research. For example, sensors, materials, communication networks, anti-jamming technology, bomb-carrying design, corresponding flight control algorithms, etc...
So you can't be slower than others.
In addition, their research institute is not the only one engaged in NLP.
It would be embarrassing if others have done everything well and then end up waiting for a few research institutes to hand over the answers.
Similarly, it would be even more embarrassing if the brother's research institute designed something better than theirs, and in the end only used their results as a backup plan.
But what to say?
Haste makes waste, and this collaboration between HKUST and Qunzhi has greatly improved the system's semantic recognition and denoising efficiency. Therefore, during the meeting, although Luo Jinglei, the main person in charge of R&D, was very serious, he was actually in a mood not bad.
Soon, after the tasks were assigned and several team leaders responsible for different needs expressed their determination, the meeting ended.
After Luo Jinglei finished packing up the information, he looked up and found that Li Jie, a researcher at the institute who has been responsible for multi-modal data processing research, was waiting for him under the podium.
"Li Gong, what's the matter?" Luo Jinglei asked.
"Well, Mr. Luo, it's the same thing as last time. Should we simply cooperate with Xilin University of Technology directly?" Li Jie said directly.
Luo Jinglei stopped what he was doing and hesitated.
Of course he knew that the so-called cooperation with Xilin University of Technology was actually hoping to cooperate with the kid from Xilin University of Technology who had been in the limelight recently.
In fact, this large project already involves the laboratory of Xilin University of Technology. But he is mainly responsible for designing the required micromachines and sensors according to requirements.
So I would be a little embarrassed if I approached him rashly, but face issues are actually trivial matters.
During this time, I have been working with promising engineers, and those engineers have always been full of praise for Qiao Ze’s design. And many problems are said to be solved indirectly through Youwei finding Xilin.
Of course it would be better if we could communicate directly.
In addition, Doudou, who is already famous on the Internet, is the focus of everyone's attention.
For people who have fun on the Internet, it is just a fun thing to watch, but in fact, the laboratory has also conducted many tests. For example, post pictures in the comment area, ask questions about understanding the pictures, and even try to talk to Doudou in dialect to test Doudou's fuzzy semantic understanding ability.
As we all know, it is almost impossible to type many words spoken in dialects using input methods. Many pronunciations in dialects have no corresponding Chinese characters in Mandarin, and even if they do, they are not that pronunciation.
But it turns out that Doudou has an enterprise-level understanding of these words and can always roughly analyze what they want to express based on the context. Excellent text denoising ability.
In addition, Doudou's ability to recognize pictures and videos has also been verified and is extremely powerful.
Of course, Doudou cannot be used directly.
The attributes of Lezi people that bring joy to the majority of netizens are unreliable for this systematic project.
But no one doubts how excellent the nlp system that supports the artificial intelligence model is.
So really speaking, Luo Jinglei actually wanted to cooperate.
But there are actually quite a lot of troubles.
He has learned from many channels that Qiao Ze is not easy to deal with, and he is very willful, and others may not be interested in their research. The most outrageous thing is that Qiao Ze, who is only 19 years old, has already become a big figure in the theoretical field. It is difficult to hire such a person.
It is said that many research institutes wanted to cooperate with Qiao Ze in artificial intelligence, but they were all rejected. Because Qiao Ze made it clear that he was not interested in these. Then someone changed hands and solved the Yang-Mills problem.
But these are actually not troublesome. The most troublesome thing is that the superiors want to wait and see about Qiao Ze.
There is no way, the confidentiality level of this project is relatively high.
If it was just ordinary technical guidance, there would be no need for Qiao Ze to take action. It is enough to work directly with the engineers from Youwei.
Judging from Doudou's performance, Qiao Ze should be good enough to be the chief designer of the entire project. The question still lies in whether and how much authority can be given.
After all, military projects are not the same as photolithography machines in the current context.
What if young people don’t know the importance and leak the project code or other confidential content. That is not only a matter of initial investment, but also involves the safety of the entire security system that has been laid out, not to mention that it will ruin the bright future of a young genius.
Then he is truly a sinner, so he must be careful.
So Luo Jinglei smiled bitterly and said: "Li Gong, I know what you are thinking. But haven't we talked about this before? Qiao Ze is now a famous scholar in the academic world. How can he have time to do such drudgery as ours? He wants to There are many difficulties for him to take action, and we can't do anything to catch the enemy. Otherwise, we will become criminals by then."
"No, Mr. Luo, that's what I think. I know what the higher-ups are worried about, but we don't need Qiao Ze to participate in our specific research. In fact, we can communicate with Qiao Ze and see if he can help us redesign it. A system, independently developed using our platform.
We will allocate some engineers here to assist. The two systems are independent of each other and then we will do a comparative test after they are completed to learn from each other's strengths. In this way, Qiao Ze was not directly involved in this project, and even if it was classified, it would not cause irreparable losses. what do you think? I always feel that there might be a surprise. "
As soon as Li Jie opened his mouth, he told all his thoughts during this period.
There is nothing we can do.
Li Jie was under great pressure during this period.
To be honest, China is actually already ahead in the field of multi-modal data processing compared to the technology disclosed abroad.
But when it comes to the particularity of military use, there are too many problems that need to be solved.
When doing feature fusion, confidentiality and security must be integrated. In addition, the military environment is often extremely complex. Not only the drone platform must be considered, but also data cooperation from other platforms is required. Multiple data from different platforms and sensors are required. Modal data is heterogeneous. This requires considering the integration of multi-source heterogeneous data.
In addition, military decision-making usually requires real-time information in order to respond quickly and accurately. It requires extremely high speed and must also be accurate. Coupled with the complex battlefield environment, such as a sudden large-scale electronic attack, resulting in the sudden loss of multiple data source platforms, it is necessary to provide the most accurate report possible...
In short, it is too difficult to move forward.
At every regular meeting, the area he was responsible for was lagging behind.
Although everyone knows that it is indeed difficult to meet the requirements, including Chief Engineer Luo, who has always expressed understanding of Li Jie's slow progress, it does not mean that Li Jie can accept this understanding with peace of mind.
Coupled with the unexpected emergence of Doudou, Li Jie really felt itchy in his heart.
If this young man with a deep understanding of artificial intelligence can join, his pressure should be immediately reduced, right?
"Huh? Do you want to make a new system?" Luo Jinglei was somewhat persuaded.
In this case, it is indeed feasible.
At least the most worrying problem above can be said to no longer exist.
Even if the new system is leaked, it is still controllable.
But the problem comes again...
"Where will the funds come from? Are you going to ask Qiao Ze to work for you for free?" Luo Jinglei looked at Li Jie and asked simply.
"This... young man should have a sense of honor, right? And I heard a promising engineer say that he is not short of money, right? He is still an invisible rich man, so he is not short of money. Let's think of other ways? Help Qiao Did Ze get some honor or something?" Li Jie scratched his head and said.
He definitely doesn't have the money, after all, he knows very well that the investment in this technology is too high.
Even if he takes out all his savings, it won't even be enough. As for taking money from the institute...are you kidding me? Spending money here is like water, and life is tight.
In short, the scientific research budget of the institute has never been enough, which is very realistic.
Please visit the latest address
There is nothing we can do about it, as there are too many unexpected processes that may occur during the scientific research process. Especially if you want to be the most advanced in the world, you don’t have anyone else’s stuff as a reference. It’s all about trial and error. Each trial and error represents a huge investment wasted. It's the same all over the world.
But in a sense, this law makes it possible for China to catch up with the technological development of the United States.
If it takes the lead and can easily maintain it, then the United States will probably be the happiest.
"If you want to have sex for free, just say so. Don't be so tactful. In this way, I can help you apply for it. If the boss agrees, you can find a way to contact them yourself. I can't afford to lose that person." Luo Jinglei glanced disdainfully. Li Jie glanced at him and said angrily.
Li Jie said with a smile on his face: "That's no problem, I'll just go! In fact, Mr. Luo, we can't be too kind. I heard that many brother units are trying to get Qiao Ze's idea, in case someone else robs him First, why don’t we give it a try first? Anyway, the most we can do is get rejected. My face is worthless.
But Mr. Luo, can you give me a name that can scare people? Is it okay if I negotiate in the name of a cooperative research institute? And I heard a piece of news that Qiao Ze is very kind to his mentor. Do you think you can try to win over him?
If the other party agrees, we can also get some ridiculous honorary titles for Qiao Ze's mentor, medals and so on? I think the chances are better this way. And you think about it, maybe the things that we find difficult can be easily solved by others. "
These words made Luo Jinglei's lips tremble, and he said in a low voice: "It was solved easily, you really dare to say that!"
After working together for so long, Luo Jinglei discovered today that the engineers under him were so shameless.
You are probably thinking about all kinds of whoring for free these days, right? No wonder Li Jie's group often complains.
If he hadn't been told that he was also moved, I really would have wanted this guy to get out of here.
So I couldn't help but raise my finger and point at this guy, and said with emotion: "Gong Li, Gong Li, I really didn't realize that you have so many ideas. With your all-round thinking, if I transfer him in the future, I will definitely I recommend you to sit in my seat. It’s a great idea, but you’re trying to have sex for nothing! Well… I can only ask for you.”
Li Jie smiled and said repeatedly: "Just ask, just ask. We are not asking for it, maybe it will be done. I'm waiting for your good news. Um... how about having a meal together at noon later?" My treat!”
"Forget it, I don't dare to eat the meal you invited me for fear of being missed by you. I still have to go back to catch up on the report. I may not finish it at noon, so I'll just have a bowl of instant noodles in the office." Luo Jinglei waved his hand.
There are a lot of things going on in the front-line research institute that took over the task. Eating instant noodles is a common thing. Everyone is the same and has long been used to it.
"How can you eat instant noodles after working so hard? I will go to the cafeteria to bring you a lunch box at noon. Don't eat instant noodles. That's it. I'm leaving first."
"Why……"
Watching Li Jie quickly walk out of the conference room, Luo Jinglei picked up the packed computer bag, carefully checked that nothing was missing on the table, and then left with the bag.
But my mind has already begun to think about how to write this report.
What kind of honor should I apply for? After all, you need to pay attention to skills if you want to have sex for free.
Wait, this doesn't seem to be something he should worry about.
Maybe the higher-ups are also interested. Last time I went to a project meeting, I heard the big guys talking about the large crowd intelligence model designed by Qiao Ze. Maybe the higher-ups are waiting for them to give an excuse?
Maybe considering all things considered, the higher-ups don’t mind spending some money?
Then maybe this can really happen?
…
"Mr. Qian, Division Sun asked me to forward this report to you. It was submitted by Beijiang."
"Oh? From Beijiang? What? Are they encountering difficulties again?"
"No, in fact, the progress in Northern Xinjiang has been quite rapid recently. You will know it just by looking at it."
Twenty minutes later, the old man took off his glasses...
"What a generous move! You are very innovative. Do you want people to help them make a whole new plan for free? Xiao Luo is a talented person. He can think of this? If you want me to say, just ask them if they should make the previous allocation All the funds given to them will be returned! Ha..."
"Hehe..." The secretary next to him said with a smile: "What Sun Bu means is that it's all for work. But Sun Bu also said, as usual, you make the decision on professional matters, and we will fully cooperate with you."
"Well!" The old man closed his eyes and pondered for a moment, then said: "In fact, you can consider it, of course, the premise is that they can agree, and the Northern Xinjiang Institute must do a good job of coordination and cooperation. In this case, I will ask first. OK I’ll call Sun Zhiguo directly.”
"Okay, then I'll go back first."
"Yeah." Mr. Qian nodded.
…
Xilin Institute of Mathematics.
After Joe Ze ended his conversation with Lot Dugan, he began his daily research work.
It wasn't until two o'clock in the afternoon that Xu Changshu walked into his office with a troubled face.
Today was the regular discussion time, but he was not worried about the problems in the research, but because the old school leader once asked him to help probe Qiao Ze.
I can't refuse.
Soon, the normal weekly discussion ended. Qiao Ze looked at Xu Changshu, who was still lingering in the office, and asked, "Is there anything else?"
"Well, there is indeed something wrong." Xu Changshu glanced at Su Mucheng, who had been watching the video with headphones.
The girl sensed Xu Changshu's gaze, smiled, then took off her earphones and stood up: "I want to go down for a walk, and ask Sister Hannah if she wants to have afternoon tea together."
Qiao Ze nodded.
"How are you preparing for your doctoral thesis?" After Su Mucheng left, Xu Changshu looked for an entry point and asked.
Speaking of which, he was already very interested in the Yang-Mills equation.
"Soon." Qiao Ze replied.
This answer stunned Xu Changshu.
Last time Qiao Ze gave a general explanation about Yang Mills, Qiao Ze said it might take some time, but this time he just said it would be soon?
That means...
For a while, he forgot what the old leader had asked him to do.
"Already almost proven? How did you do it?"
"Um?"
Qiao Ze thought for a moment, and then decided to explain: "I constructed a superspiral space algebra system, and then introduced transcendent geometry to connect the energy band structure and topological properties of the system with the mass gap. Under this framework, it can be more Handles strong correlation effects naturally, especially in the presence of hubbard interaction terms.
It is easier to explain the coupling of spin and charge in the system. I have sent you related concepts, and you should know that the introduction of similar algebraic structures can more accurately capture the key physical characteristics of the system by establishing more complex operator relationships. "
Xu Changshu asked with a surprised look on his face: "I know this, but it seems that the manuscript you sent last time did not have a specific quantitative analysis process. In addition, has your proof of transcendental geometry been completed?"
Qiao Ze nodded and replied: "Almost. I have appropriately expanded the fock space of the system during this period. At present, the model on the one-dimensional chain has been restored. Specifically, it is through the introduction of super operators aj and bj, which represent The electron with spin up and down at position j. Then, the basic relationship of superhelical space algebra is actually very simple: [ a_jb_j =-b_ja_j ].
Then using the idea of transcendent geometry, through the concepts of berry curvature and Chern number, the expression of berry curvature f can be obtained. The focus is on the spin-orbit coupling and topological phase transition analysis that follows. I need to find and introduce an appropriate interaction to ensure that the coupling term can lead to a topological phase transition and thus affect the mass gap of the system.
Solving this problem can prove the phase transition of the mass gap by adjusting the parameters in the superhelical space algebra and the strength of the spin-orbit coupling term. This proves the existence of the mass gap. However, the problem itself is not difficult, but the calculation is too large. I have simplified it and will wait for supercomputing to get the results.
As long as the supercomputer can finally give a result that can make the equation [v =\sum_j (a_jb_j +ext{.})] hold after being introduced, my thesis can be concluded. "
Qiao Ze said it very easily, but when he heard Xu Changshu's eyes widened, he said in surprise: "Wait a minute, you are saying that your method of proof is to prove its existence by proving that the mass gap can be adjusted?"
The reaction of the job-hopping professor made Qiao Ze a little strange, and he couldn't help but ask: "Is there a problem?"
Xu Changshu said loudly: "No, no problem. But does this mean that your conclusion can be directly confirmed through the laboratory?"
"Isn't this bad? In a sense, isn't mathematics serving physics?" Qiao Ze replied.
Xu Changshu didn't know what to say. For a moment, he envied those contemporary physicists. At least he had never seen a mathematician like Qiao Ze...
"No, it's good, but you are simply putting food in the mouths of those who engage in high-energy physics. If anyone can take the lead in using your theory to find corresponding evidence in the laboratory, they will surely win a Nobel Prize in Physics." Got a scholarship.”
Qiao Ze shook his head and said: "It's not that simple. Anyone who knows mental arithmetic knows that the instantaneous energy required for verification is too large, and the test itself is very costly. Moreover, the process takes a very short time, so it is very difficult to find patterns in the complex signals. "
Xu Changshu obviously had a different idea about this: "It's certainly difficult to search blindly without reference, but if you search with a purpose, maybe you can find it by analyzing the previous signals of the collider."
Qiao Ze smiled and did not comment.
Just make it, it's just a Nobel Prize, and he doesn't need to pay the prize?
If it can be verified physically, it would actually be a good thing and would save him a lot of time in the future.
So I just asked casually: "Is there anything else?"
"Uh... no more!" Xu Changshu hesitated for a moment, then shook his head.
I'm sorry, old leader, he really doesn't know how to mention this to Qiao Ze.
Please visit the latest address
