Dear Jiankang,

My opinion is negative.

That would probably occur in a few early products, which would eventually fail/transform due to the computational expenses required, and pricewise.

In the future (2040-2050), there will be cloud centers managed by the Giants (Alphabet of that time, Amazon, Huawei, etc.). Then all personal devices, tablets, and PCs will be connected to supercomputing AI centers through Internet and will do all intelligent processings for every and each device/PC owners using supercomputing technologies of that time (which may be DNA supercomputing).

Therefore, this idea of running AI in OS/kernel layers of present products will be solely short-term, if become feasible and so, which I highly doubt that, due to being not cost-efficient.

Perhaps for some privacy and security reasons, some of such systems of such high prices would only be used by banks, govs, etc., but I highly doubt that every and each PC and device owner carry some sort of AI agents in hardware layers.

I may be wrong though.

Best regards,

J.

Applying AI in PHY layer for Resource Allocation ?

There is a fact that artificial intelligence will cause; All of the work that is still being done and predictable by people will be done by computers. With the development of artificial intelligence, 50% of the works will be lost in our world in 15-20 years.

Let's give a few examples of some of the different sectors;

Health. Patients will explain their symptoms to the artificial intelligence at the other end of the phone, using their cameras will show where the artificial intelligence wants to see. The artificial intelligence system will also advise patients to take a painkiller or recommend a doctor as soon as possible.

Artificial intelligence will read and interpret medical images. Thus, doctors will speed up and facilitate their work.

Artificial intelligence, using the data will help doctors to diagnose disease. The patient may have hundreds of symptoms leading to symptoms. Even the best doctor can't remember everything.

Education. Students will do lessons with artificial intelligence. Artificial intelligence will be able to get to know the students very well, follow their learning situations, and move on to students at the appropriate speed and subjects. Training will be truly personalized.

Finance. Exchange traders will transfer their financial investment decisions such as stock buying and selling to artificial intelligence.

In banks (and similar sectors) back-office workers will do their work. Today, Bloomberg company is printing financial reports of public companies to robots.

Transportation. Driverless (artificial intelligence) cars and trucks, buses, minibusses, minibus drivers will not need. at least 4 million truck and lorry drivers in Turkey available.

Entertainment. Travel agencies will be destroyed by service and service-based service sectors. This kind of work will make artificial intelligence.

Other. The professions such as waitresses, office management, cleaning, and research will be done by artificial intelligence. For example, we will listen to our meetings with artificial intelligence and provide us with useful information on the internet.

Dear Jiankang,

My opinion is negative.

That would probably occur in a few early products, which would eventually fail/transform due to the computational expenses required, and pricewise.

In the future (2040-2050), there will be cloud centers managed by the Giants (Alphabet of that time, Amazon, Huawei, etc.). Then all personal devices, tablets, and PCs will be connected to supercomputing AI centers through Internet and will do all intelligent processings for every and each device/PC owners using supercomputing technologies of that time (which may be DNA supercomputing).

Therefore, this idea of running AI in OS/kernel layers of present products will be solely short-term, if become feasible and so, which I highly doubt that, due to being not cost-efficient.

Perhaps for some privacy and security reasons, some of such systems of such high prices would only be used by banks, govs, etc., but I highly doubt that every and each PC and device owner carry some sort of AI agents in hardware layers.

I may be wrong though.

Best regards,

J.

I partially agree with J. Rafiee

"In the future (2040-2050), there will be cloud centers managed by the Giants (Alphabet of that time, Amazon, Huawei, etc.). Then all personal devices, tablets, and PCs will be connected to supercomputing AI centers through Internet and will do all intelligent processings for every and each device/PC owners using supercomputing technologies of that time (which may be DNA supercomputing)." ㅡI AGREE WITH YOU.

Btw, I think that the system would work only after individual device can be controlled by AI-integrated operation systems.

I think it will sucessfully be applied at a point when human predictions prove ambiguous.

Hello,

The actual application of AI today is for classification that helps in education, health, finance etc. Since network infrastructure are well defined systems there is no need today or in the near future to apply AI everywhere in the system. In a matter of fact, Application Defined Network concept is being developed to create virtual networks and optimize the operation of network infrastructures where partial AI can be applied to provide a smart network performance.

Thanks,

Your question involves many other questions. And what do you want to get at these levels by applying AI in the PHY layer, MAC layer, network layer?

Well, then, the same routing has every reason for redistributing traffic using AI. The same work at the MAC level. But rather, the work on the last two levels, except the network level, will rather be rigidly algorithmized, than non-standard solutions will suddenly emerge. Therefore, my answer is the use of AI in PHY layer, MAC layer, no. The use of AI network layer is even very possible ..