After we started to feel that the world of smartphones is heading towards a relative stagnation in the field of revolutionary matters and that new phones only come with improvements in performance, camera and design, it came in 2017 to reveal a new trend as companies began to reveal what they call “AI processors to support artificial intelligence and machine learning” and for each A company has its different way of using and developing this processor, but what are these processors that companies can be proud of?

AI processor in smartphones

Huawei announced at IFA 2017 in the German capital Berlin the launch of the new Kirin 970 processor that it used in its Mate 10 device, describing it as the first processor of its kind as it was equipped with a neural processing unit called "NPU" or Network Processor Unit and also described it as "true artificial intelligence." The processor is capable of parallel and immediate processing in real time, and is also able to perceive cognitive tasks such as understanding, forecasting, learning, the ability to sort data, analyze it, classify it, and much more like the human mind, and these processors use lower energy than other processors.


Then Apple announced the A11 processor with a Neural engine in its iPhone 8, iPhone 8 Plus and iPhone X devices, and the company said that it was designed specifically for "machine learning" in addition to many other uses as well.


A few days ago, Qualcomm announced the Snapdragon 845 processor. In fact, it is a powerful processor that focuses heavily on photography and makes it more realistic. It includes an Adreno 630 graphics processor, and is also equipped with the Spectra 280 image signal processor, which allows shooting "Ultra HD" premium video at a rate of 60 FPS, this processor exceeds a score of 100 in DxOMark Rating .


What is an AI processor

In future days you'll hear the term AI processor as often as you'd heard in the past about CPU processor. The AI ​​processor refers to multi-feature computing, which in turn refers to systems that use multiple types of processors, each with specific functions aimed at strong performance and energy savings, and the truth is that this idea is not new as many of the chips now present are based on this system.

In the past three years, CPU processors in smartphones have relied on ARM's big.LITTLE technology. And for those who do not know big.LITTLE technology is a technology provided by ARM to save energy while maintaining the power of the processor by switching the task command process between the processor cores, meaning that the processor will be divided into two parts, the first section will be dedicated to the powerful tasks, and the second section It will have tasks that do not need a powerful processor to perform. And this technology was used in the past in processors phones Samsung Galaxy S4 and Huawei's Mate 8 and Honor 6.

This year, AI processors took the previous concept a step further by adding a component dedicated to carrying out machine learning tasks. For example, the Snapdragon 845 processor can take advantage of the Digital Signal Processor (DSP) to handle time-consuming tasks that require frequent mathematical operations, such as listening to important words such as "Hey Siri" on the iPhone, or "OK ​​Google" on Android devices. .

In addition, Gary Brotman, Director of Qualcomm Products Department and Head of the Department of Artificial Intelligence and Machine Learning at the same company, said, "Many activities, including image recognition, are better managed through the graphics processor integrated with the new Qualcomm processors."


Meanwhile, Apple's A11 processor uses a neural engine built into its graphics processor to increase the speed of real-time recognition of facial and animated faces and some third-party applications.

This means when you use the previous things on the iPhone X, the A11 processor gives a signal to the neural engine to perform the necessary calculations and draw a face map to verify you or to recognize your different facial expressions while using the Animoji or other applications.


And in the Kirin 970 processor, its NPU scans the image and translates the words in it through the Microsoft Translator application, which is the perfect application to do such things.

Huawei said that the combination of heterogeneous or multi-feature computing in the AI ​​processor increases the performance of the rest of the components on its chips, so the tasks of this processor go far beyond being a neuroprocessor dedicated to specific tasks only.


Among the features of the AI ​​processor

◉ The new architecture of this processor, which primarily means machine learning and artificial intelligence, which was previously based on cloud computing, can now be implemented on the device more efficiently.

◉ Among the advantages, when you develop applications or programs based on artificial intelligence, it will be easy to develop and test on the phone itself without the need to use external services over the Internet.

◉ And by using these processors on your phone, you can do more than one thing at one time, such as instant translation by scanning texts with your phone's camera instead of sending them to the Internet.

These features are better in terms of privacy, which means that it is possible, even in a small percentage, that this data is penetrated during the transmission and reception operations, but with these chips you can perform these operations without the need to send and receive any data over the Internet.

◉ Among the advantages also: saving energy, we all warn the most caution when using applications, so you find most of us look every moment at the battery sign and at how much the percentage has reached, and it gets worse if you are on a travel and do not take the charger with you, so you have to do the low power mode and switch off The phone data, Wi-Fi, and Bluetooth make the phone's lighting to less than half, and if it requires you to turn off the phone to save energy, I will. It is our right to use all the capabilities of the phone without disabling any feature, so why are we forced to disable it to save the battery?

It seems that the matter is different this time, as they say, as with the development of these new processors, the optimal development, and their support in systems and applications, the intensive use will not require much effort from the AI ​​processor, unlike other processors used frequently today, as it does its best to run applications, which consumes a lot of energy the battery.


Important clarification

To be more clear, this technology is still in its early stages and we have not seen its noticeable benefits in the real world yet. Also, this technology has not been fully supported by the developers, and perhaps the reason is that it is not widely supported in phones, it is only in the iPhone 8 / X and the Huawei Mate 10 phone, and other companies such as Samsung, HTC, Sony and others are waiting for the Snapdragon 845 processor to support it.

Even in companies that use this technology, they have only used it in specific cases, such as Apple's Face ID and Animoji. We just have to wait until we see the results of the inter-company races and release the true benefits of these treatments.

What do you think about the idea of ​​AI processors? Do you expect that it will be utilized in a real and expanded future? Let us know in the comments

Source:

engadget

Related articles