When I am reborn, I just want to be a top student

In the conference room, after Wang Donglai finished his polite remarks, he got to the point of the meeting.

Wudaokou naturally has certain expectations and prospects when it established such a research team.

Faced with this brand new scientific research project, not only Yao Qizhi had little exposure to it, but even the other invited experts had no exposure to it before.

You have to know that the domestic Internet industry claims to be innovative, but it is just moving some mature industries from developed countries over.

This is true for social networking, games, food delivery, portals, etc.

But the real core of the Internet doesn't care about this at all, and it's not about this.

As early as 2012, Google's Artificial Intelligence Laboratory proposed enabling machines to learn language and grammar.

What was the domestic Internet doing at that time?
Sit down!

At the head of the conference room, Wang Donglai bowed and then stood up.

As he stood up straight, a unique aura emanated from him.

He has the enthusiasm of a young man, the confidence of a scientific researcher, and the aura of a business tycoon.

"I don't know if anyone here has heard of GPT, which was presented at the Bald Eagle Artificial Intelligence Academic Annual Conference last February. Let me briefly introduce it here."

"In 2012, Google's AI lab proposed to let machines learn language and grammar. The following year, it released a language model based on GPT, which learns language by encoding the input."

"In October of the same year, Google started a pilot project for GPT and made improvements to enable it to learn language models without requiring a large amount of training data."

"So, our goal is actually very simple, which is to benchmark GPT and start with a large artificial model that is unique to China."

"The module function that needs to be completed in the first phase is to be able to communicate like a human, and even complete functions such as writing emails, creating copywriting, and mathematical calculations."

When Wang Dong said this, the professors and researchers sitting below became a little agitated.

It was really that the requirements put forward by Wang Donglai were too high and completely exceeded everyone's imagination.

Communicate like a human, write emails, and create copy.

According to Wang Donglai, this function is only in the first stage.

Just by listening to what Wang Donglai said, everyone quickly thought about it and came to a conclusion.

That's simply impossible.

They simply can't do this with their abilities.

Although this is their first time to come into contact with artificial intelligence, they still have some understanding of computers, and they are also well aware of their own abilities.

If they could really do what Wang Donglai said, they wouldn't be like this now.

Immediately, a middle-aged man wearing glasses interrupted Wang Donglai.

"Professor Wang, you just mentioned that the module functions to be completed in the first phase are to be able to write emails, create copy, etc., but as far as I know, such technology has not been completed even in Bald Eagle."

"I have also heard about Google's GPT project. According to communication with some overseas friends, GPT was first proposed by the Opanai research team. It mainly uses neural networks and deep learning to enable machines to acquire powerful natural language processing capabilities. Only in this way can the model have tasks such as generating text, answering questions and performing language translation."

"Let's not talk about whether this can be done. Natural language processing alone is already an extremely difficult subject, not to mention that artificial intelligence involves many fields such as computer science, psychology, and mathematics. The research fields include but are not limited to robotics, image recognition, expert systems, etc."

"Even the artificial model we are currently working on is not a simple topic. It is not accurate to call it an artificial model. It should be called a large language model, also known as an LLM. It is a model that uses deep learning algorithms to process natural language. It can identify, summarize, translate, predict and generate text and other content based on massive amounts of data."

"So, in this case, our research team has a heavy task to accomplish. I'm not trying to undermine Professor Wang and everyone else's confidence, but I'm actually not very optimistic about the current team."

Before Wang Donglai could speak, Yao Qizhi asked, "Engineer Liu, what do you think? As long as what you say makes sense, you can speak freely in this team!"

The middle-aged researcher called Engineer Liu glanced at Yao Qizhi, then looked at Wang Donglai who was silent, and then continued: "It's very simple, that is to add more manpower!"

Hearing this, Wang Donglai laughed.

"Engineer Liu, I didn't expect you to have such a deep understanding of artificial intelligence. It seems that this project has a greater chance of being completed."

"As for the question Mr. Liu just raised, I have actually already considered it."

"First of all, our project this time is a large-scale language model. I believe Mr. Liu has no problem with this. Large-scale language models are nothing more than the use of machine deep learning to complete. We can develop a language model based on recurrent neural networks to generate coherent text, etc."

"The second is natural language processing, which Mr. Liu mentioned. I also have a solution for this. That is to use the Transformer decoder to complete a language model without labeled big data as a pre-training model, and then do fine-tuning on the subtasks to obtain classifiers for different tasks, so as to achieve natural language reasoning, question answering and common sense reasoning, semantic similarity recognition, classification, etc."

"Finally, there are the areas of algorithms, image recognition, and expert systems that Mr. Liu mentioned, and I am also prepared for them."

"I'm afraid Mr. Liu doesn't know me well enough. He doesn't know that my first SCI paper was about algorithms. I think I'm not inferior to others in this area."

Wang Donglai spoke clearly and in a calm tone, and when he said this, there was an unspoken aura of oppression.

At this moment, everyone's impression of Wang Donglai changed drastically.

How gentle and polite Wang Donglai had behaved just now, and what a contrast there is now.

There was not a single curse word, and no change in tone.

However, his words were filled with strong self-confidence and dominance, which made everyone involuntarily look at Wang Donglai with admiration.

Wang Donglai's performance did not surprise everyone. It is normal for him to achieve such achievements at this age and have such performance.

Many people turned their eyes to Mr. Liu, wanting to see if he would say anything else.

Unexpectedly, after receiving such an answer from Wang Donglai, a thoughtful look appeared on Mr. Liu's face.

"Professor Wang, you just mentioned developing a language model based on recurrent neural networks and natural language processing. It sounds quite feasible. Can you tell us more about it? I feel that if these technologies are solved, our research in the field of artificial intelligence will probably be at the forefront of the world."

“Even if this is just one aspect of artificial intelligence, an application product, as long as a breakthrough is achieved in this field, the impact will be huge, enough to cause a huge shock in society.”

The more he spoke, the brighter Mr. Liu's eyes became, and his expression became more excited.

Seeing this, Wang Donglai had a clear understanding of Mr. Liu in his mind.

A true researcher!

He interrupted me just now, not because he had any opinion on me, but because he couldn't stand seeing someone talking nonsense, so he interrupted me. But after I said some ideas, Engineer Liu was able to react in such a short time, which proved that Engineer Liu had excellent vision and profound knowledge and technical reserves.

Wang Donglai was very pleased that his research team had such a capable researcher. He smiled and said, "Liu, artificial intelligence has a wide range of applications and involves a wide variety of issues."

"For now, it is enough to be able to develop a domestic general model of artificial intelligence chat!"

Wang Donglai didn't say much, and he wouldn't mention all the key points here.

As he said, some core algorithms will be completed by him, because others cannot do it firstly and secondly because of the need for confidentiality.

Being half a step ahead of the times is a genius, being one step ahead is a monster, but if you are ahead by several hundred meters, that would be a disaster.

When Mr. Liu heard what Wang Donglai said, he opened his mouth but said nothing. He nodded and fell silent.

Seeing this, Yao Qizhi took over the topic.

"I believe everyone has seen the details of this research project before. This is the first of its kind in China, so we have a pioneering responsibility."

"As for this research topic, Professor Wang is the core. I am not afraid of being laughed at when I say this. I feel that Professor Wang actually knows more about artificial intelligence than I do and has more knowledge reserves. The discussion between Professor Wang and Engineer Liu just now proved this point."

"What's more, this project is also funded by Galaxy Technology, and the results and patents are exclusively owned by Galaxy Technology."

“Although I am the leader of the next research work, I hope that you can show your true strength in the research process and listen to Professor Wang’s advice when you encounter problems.”

Yao Qizhi picked up the microphone and said this to everyone in great earnest.

As Yao Qizhi said this, the expressions of everyone present suddenly became serious.

In fact, with the previous communication between Wang Donglai and Engineer Liu, everyone had already understood it.

Wang Donglai’s accomplishments in this area are actually quite high, and it can even be said that he has surpassed most of the people present here.

Some people who are interested in this have even listed in detail that, apart from Wang Donglai's achievements in mathematics, his performance in computer science is also quite good. The fact that Wudaokou was able to cooperate with Galaxy Technology so readily on this project is probably because of Wang Donglai.

Having such a genius join the team will only bring benefits and no harm to the artificial model project.

Everyone can understand this point.

So, after Yao Qizhi finished speaking, everyone nodded.

"Don't worry, Dean Yao. Professor Wang's achievements are here. We will certainly not turn a blind eye to them."

"If artificial intelligence technology can achieve a breakthrough, it will be of great significance to both the country and the academic community. We will naturally not hide our shortcomings and will definitely go all out!"

"It just so happens that I don't have any projects to do, so I can focus more on this project next time."

"What a coincidence! Several of the projects I'm working on have reached the final stage and I can just hand them over to the students I'm supervising. This is a good opportunity for me to devote myself to this project!"

"..."

All of a sudden, everyone in the conference room expressed their opinions.

Wang Donglai was not too surprised by everyone's performance.

University professors actually have quite a few research topics on hand, especially those in Beijing.

For a project, most of the work is actually handed over to the students below, and doctoral students and graduate students are the main workers.

They have to do all the dirty work, and the professors only give guidance on key points.

This is not to say that these professors are deliberately exploiting students, but because they have too many research projects on hand and this is the best option when they are unable to handle them all.

There is also the consideration of training the students' abilities.

Of course, there are many kinds of people in the world, and naturally there will be unscrupulous professors who exploit the labor of students. This is inevitable.

Even Tangdu Jiaotong University has it, but at the suggestion of Wang Donglai, Zheng Hua and Xu Songyao imposed strict restrictions and supervision on this aspect in an extremely tough manner, which made the academic atmosphere of Tangdu Jiaotong University much better.

but now.

These university professors and research institute researchers have put aside other research projects they were working on and handed them over to their graduate students and doctoral students, and have focused their main energy on artificial intelligence and big data.

The reason is very simple!
Artificial intelligence big data models are even more valuable.

At the same time, the feasibility is also higher.

Once the project is completed, it will be a huge benefit for them.

It is helpful whether it is publishing papers or evaluating professional titles.

There are three main challenges in the research and development of large artificial models, namely data, neural network architecture and hardware.

In terms of data, because a general large model is required, a large amount of data is required. Fortunately, China has advantages in this regard, so the problem is not too big.

As for the neural network architecture, with the help of big boss Yao Qizhi and many university professors, and finally Wang Donglai, there is naturally no problem.

Finally, there is the hardware part.

Although we can use the "Exploration 1000" supercomputer of Wudaokou, it is not a long-term solution.

In order to develop in the future, you must have your own machine.

This requires high-performance CPUs and GPUs, which need to be imported from abroad and are subject to control.

Even so, I have no choice but to buy it.

At present, the domestic demand cannot be met at all.

When Wang Donglai proposed a large artificial intelligence model, he had actually thought of this problem.

buy!
There is only this one choice. Before the West imposes sanctions and China catches up, we can only buy these things. (End of this chapter)