Chapter 1079 Intelligent Type II AI~
A UN General Assembly ended in vain.
For China, this was expected.
The position of the Five Constant Persons is not in vain. The United Nations is the United Nations of the Five Constant Persons, not the Five Constant Persons of the United Nations.
Their purpose will be achieved by targeting the relevant meetings back to the Security Council for internal resolution, and the rest can be slowly "discussed", and they are not in a hurry anyway.
After all, they have a veto power for the vote of the Security Council. As long as they do not agree, this discussion on the space-based electromagnetic rail gun will only be shelved indefinitely.
Of course, even if the United Nations finally convened the United Nations General Assembly under the temptation of the United States and finally passed the vote, it would not have much significance for them.
Because of the non-Security Council United Nations vote, the United Nations does not have the right to enforce it.
Unless all countries unite and do not make diplomatic and trade with them.
But this is obviously impossible.
However, the United States will definitely not give up on space-based electromagnetic rail guns.
Next, they will either continue to find ways to win over other countries such as the EU and Tsarist Russia to discuss sanctions or targeting them, or they will choose to develop their own outer space weapons.
But no matter which approach, there is nothing to worry about and fear for them.
Many times, if one step ahead, it is a gap that will never be able to catch up in the rest of one’s life.
What's more, the space-based electromagnetic railgun is far from their trump card, and the aerospace engines and small fusion reactors are!
With these two technologies, they have enough ability to assemble a large fleet in outer space whenever they need it.
Compared with the "ordinary" space-based electromagnetic railgun, the deterrence of the space fleet is the real "invincible"!
While the United Nations is holding a meeting, on the other side.
China, Jinling.
In the villa at the foot of the Purple Gold Mountain, Xu Chuan was flipping through the simulation calculation program optimized by the AI academic assistant with interest.
Although he is not a professional programmer, programming also plays a crucial role in mathematical modeling.
He is not unfamiliar with modeling and code.
Especially when it comes to those who transform real-world problems into mathematical equations, and selecting appropriate model types, such as linear models, nonlinear models, differential equations, random models and other fields, they have their own unique insights.
Otherwise, when he first studied controlled nuclear fusion technology, it would not be so easy for him to quickly apply the NS equation to model and control the turbulence of high-temperature plasma in the reactor.
"It's interesting. Here I adjusted the numerical structure of the analysis method, and changed the calculation method of data analysis and calculation, saving a lot of calculations."
"The fuzzy algorithm is used here."
Looking through the codes marked by Xiao Ling, Xu Chuan's eyes were filled with a thoughtful look.
From the perspective of being optimized, it has made a lot of changes to Warp tensors, which has greatly improved data movement, tensor core MMA instructions and CUDA core performance.
In addition, it also makes some changes to block scheduler and dense packet functions.
It seems that behind Xiao Ling, that is, Chuanhai Network Technology has a large training database, and the training of these things cannot be done with simple modifications.
After confirming that the optimized mathematical model was fine, Xu Chuan downloaded it and immediately sent a message to Zheng Hai.
"Professor, come to me."
After a while, Zheng Hai rushed over, Xu Chuan nodded and handed over the storage USB drive in his hand.
"Send it to the research institute and find Zhou Ji from the Information Institute to arrange the supercomputer and run the models and related data."
"Okay."
Nodding, Zheng Hai took the USB flash drive, put it in his arms and asked habitually: "Is there anything else? Professor."
Xu Chuan thought for a while and said, "Let me arrange another trip to Sichuan Capital. If it's time, it's Friday."
Since my mother-in-law suggested that she wanted to meet him, it was naturally impossible for the old man to wait too long.
Today is already Wednesday, and tomorrow he will arrange the work of the research institute and CRHPC institutions. Friday is just right.
After Zheng Hai left, Xu Chuan stretched and stood up to walk towards the living room.
On the coffee table, a few blooming tulips and colorful lilies were inserted in a long-necked transparent vase, blooming in a staggered height.
In addition to the flowers on the coffee table, several pots of succulents and plants are also placed beside the window of the living room, adding a bit of vitality and homely atmosphere to the originally cold living room.
This was naturally brought by Liu Jiaxin.
After all, with his personality, it is impossible to make such floral things in the room.
After fiddling with the lily petals on the coffee table, Xu Chuan glanced at Liu Jiaxin and the housekeeper who were busy in the kitchen and walked over.
"Are you hungry? You have to wait for a while for dinner."
In the kitchen, Liu Jiaxin, who was helping the housekeeper pick and wash vegetables, looked up and couldn't help but smile on her face.
Xu Chuan smiled and shook his head, saying, "No, I'm just coming out to see."
Thinking of something, he asked curiously: "By the way, how is the progress of artificial intelligence AI research on the company? When I was just conducting research, Xiaoling was able to automatically optimize the code?"
Hearing this question, Liu Jiaxin thought for a while and said, "At present, what is probably the mid-stage of the Agent Type II?"
"Agent Type II mid-stage? Is it evolution so fast?" Xu Chuan asked in surprise.
The mid-term of the Intelligent Type II is a phased evaluation of AI intelligence by Chuanhai Network Technology.
Overall, they divide artificial intelligence into five major stages, namely chat intelligence, reasoning intelligence, intelligent body, innovative body, and complete body.
These five stages correspond to the intelligence of artificial intelligence.
Each stage can be further subdivided into three models, Type I, Type II, and Type III according to the complexity of artificial intelligence.
Among them, the AI system can conduct basic dialogue and communication in the chat intelligence stage, show basic understanding of natural language, and respond to various prompts and questions. This level of artificial intelligence is mainly used in scenarios such as intelligent customer service, intelligent assistant, etc., providing basic information consultation and interactive services.
For example, chatbots in daily life can answer user questions, provide information, and even engage in simple interactions.
The reasoning intelligent AI system can solve complex problems with the proficiency of ordinary human experts, marking its upgrading from simply imitating human behavior to showing real intelligence.
At this level, AI systems can be applied to areas where complex reasoning and decision-making are required.
For example, medical diagnosis, financial risk assessment, etc.
Inference intelligent AI can deeply analyze complex data, understand the patterns in it, and provide accurate solutions based on logic and data.
The AI system of agents can undertake complex tasks, make decisions, adapt to changing environments, and act independently without human supervision.
At this stage, AI not only has the ability to reason, but also can independently perform various complex operation tasks.
Can be applied to autonomous driving, robot control and other scenarios,
For example, in the field of smart homes, the intelligent body can intelligently adjust the lighting, temperature and other home environments according to personal habits; in the field of autonomous driving, the intelligent body can ensure the safety and comfortable driving of the vehicle, etc.
If it is the initial academic AI assistant, "Qi Ling", its level is in the stage of reasoning intelligence II model.
You can use corresponding instructions to collect, classify, organize various data, etc.
The newest ⊥Xiao⊥Speak⊥Send⊥Six⊥9⊥Book⊥⊥⊥⊥Back⊥⊥⊥⊥!
After fully integrating into the complete artificial intelligence of Chuanhai Technology, Xiaoling has the ability to control smart home appliances, autonomous driving, data diagnosis, program code writing and other aspects.
According to evaluation, at that time it should be in the transition stage of the reasoning intelligence type III stage to agent AI, which is very close to agent AI, but has not yet reached it.
And this data was almost half a year ago.
In half a year, the evolution from the reasoning intelligence type III to the agent type II was indeed a bit beyond his expectations.
Originally, he thought it would take at least one year or even longer for Xiaoling to evolve to this stage.
Unexpectedly, it was done in just half a year.
Liu Jiaxin nodded and smiled, "I just made a breakthrough not long ago. The weight layer of the deep neural network in the Dazheng integer factor decomposition paper on Dazheng Group II has been applied to the intelligent AI matrix decomposition structure."
"This reduces the total number of parameters of the training model and also speeds up Qiling's training to a certain extent."
"Especially, there have been considerable breakthroughs in SVM classification construction, integrated learning module, large sparse matrix problems and fuzzy algorithms, and it is possible to solve the feature combination problem under large-scale sparse data through factor decomposition machines."
Upon hearing this, Xu Chuan suddenly understood.
It would not be surprising if it was done based on whether Taisho integer factor decomposition has some theories in the polynomial algorithm.
"P=NP?" is a central question in computing theory, although whether or not Taisho integer factorization has a polynomial algorithm is only part of this millennium puzzle.
However, it has a significant impact on multiple computer algorithm fields such as distributed computing and parallel processing algorithms, single-source shortest path algorithms, integer factorization, etc.
It has been a year and a half since Liu Jiaxin solved this problem in 2023, and it is almost time to apply it to the fields of computers and artificial intelligence.
After all, the transformation of mathematical achievements into research in the field of computer is one of the easiest areas, and there are also scholars who have proved this problem that are developing in the field of computers, and this is the reason why Xiaoling can evolve rapidly.
Unlike the mathematical problems he studied, except for the mathematical conjecture that is highly applicable in itself, the other two problems, whether it is the Hodge conjecture or the Yang Mills existence and mass gap problems, even though it has been solved for several years, there is still no movement in the application field.
The easiest problem in the field of mathematics needs to be transformed into application results, probably the P=NP problem.
But the P=NP problem is not that easy to solve.
At least for now, both he and Liu Jiaxin, who solved the problem of whether Taisho integer factor decomposition has polynomial algorithms, believe that this problem may not be solved.
If God does not roll dice, it is almost impossible for humans to find a polynomial algorithm to verify that NP-class problems belong to Class P.
But for computer development, P=NP? The possibility that the puzzle is unsolved does not mean that it cannot continue to develop.
On the contrary, every study of this problem, even a little breakthrough may play a huge role in promoting the algorithms of computers and artificial intelligence.
Just like in the 1970s, people made revolutionary breakthroughs in cryptography based on NP complete theory and established a public key cryptography system.
Then in the 1980s, the research on complete problems made vertical breakthroughs, and a profound portrayal relationship was found among many computing models that seemed unrelated on the surface.
This breakthrough has greatly stimulated the new wave of research on approximate algorithms in the algorithm community.
In addition, there are research on quantum computers and proposition proof systems in the 1990s, and theoretical research on parameter complexity after the 21st century.
Almost every time P=NP? Breakthroughs and research on mathematical achievements in conjecture will cause a sensation in the field of computers.
If "Qing Ling" continues to evolve, it will be not far from the fourth stage of "Innovator AI".
The intelligent AI system at this stage has the ability to imitate human creativity, break through the limitations of thinking, and propose refreshing innovative ideas.
Simply put, it has been able to propose solutions or ideas for a certain problem based on a huge database that humans have not thought of before. To a certain extent, it is weak, creative or weak, original.
At this level, artificial intelligence AI can assist humans in invention and creation, and promote innovation and progress in multiple fields.
For example, in many fields such as scientific research, artistic creation and technological development, innovators can exert their unique value and provide new ideas and methods.
Even you can turn P=NP? I guess it is left to deduce such problems, try to solve them, etc.
Although it may not necessarily be able to do it, in theory, it is already fully able to think about this highly complex problem.
As for the final organizer AI level.
In theory, it is already a complete artificial intelligence.
At this level, it not only has strategic thinking, but also has the high efficiency and adaptability needed to achieve organizational goals and is able to manage complex systems.
For example, they can flexibly coordinate multiple intelligent AIs, allocate tasks reasonably, monitor task progress in real time, and make quick adjustments based on actual conditions, etc.
In theory, you can give a large company or an entire country to it for management.
It is sufficient to develop ‘perfect’ and ‘fair’ plans to maintain the operation of the company and the country.
PS: There will be some refills tonight, but it will be relatively late. Everyone will watch it again in the morning.