HomeNewsHow AI features in smartphones reduce their dependence on the cloud

How AI features in smartphones reduce their dependence on the cloud

New phones are being launched with features enabled by artificial intelligence (AI). The latest of those was Google’s Flagship phone Google Pixel 9. Samsung's Galaxy S24 phone, due out in early 2024, also features a spread of AI-powered photo editing features.

The hidden story behind such devices is how corporations have managed to maneuver the processing required for these AI capabilities from the cloud to the device within the palm of your hand.

In the Google Pixel 9 phone, a feature called Magic Editor allows users to “reinterpret” their photos use generative AI. What does means in practice is the power to reposition the topic within the photo, remove one other person from the background, or change the grey sky to a blue one. This is finished by providing appropriate prompts and the app does the remainder.

The phone's generative AI features also permit you to add people or objects to your images by typing a text prompt.

Of course, users have all the time been capable of do that with photo editing software, but making the result look natural and never prefer it's obviously been edited requires some skill. Magic Editor guarantees to make use of AI to perform these complex photo edits with “easy and intuitive actions.”

Another feature called “Add Me” allows users to take a bunch photo without having at hand the phone to a stranger. The owner of the phone simply takes a photograph of the group, then gives it to a friend and goes to the identical place they simply took a photograph of. The phone then stitches the 2 shots together.

Another feature called “Best Take” will be used to pick out the most effective elements from a series of very similar images and mix all of them into one image. Google's chatbot technology powers a digital assistant and other features on the phone.

The functions of telephones have evolved significantly since then. the primary digital telephones; or when phones began having their very own built-in cameras.

2001 BBC Archive: First smartphone (cameras on phones)

To the sting

Until now, the processing required for such AI-based functions has been too sophisticated to run on a tool reminiscent of a phone, so it’s as a substitute outsourced to online cloud services powered by large, powerful computer servers.

However, corporations are increasingly recognizing the necessity to have much of their data processing done on customers' devices, which could potentially give consumers more control.

Considerable amounts of AI computing processes are transferred to the so-called “Edge”. The edge describes typical consumer devices reminiscent of phones with reduced processing power.

The difference between how cloud-based and edge-based AI works:


John Chiverton, Provided by the creator (no reuse)

To do that, the ability consumption for processing should be reduced. Companies have achieved this with special microprocessors which might be specifically tailored to AI-based processes.

For example, Google's tensor AI processors, called Tensor Processing Units (TPUs), seem like central to the capabilities of their Pixel phones. The edge-based processors are capable of efficiently apply AI models to data collected or stored on mobile devices. with the assistance of special software.

These TPUs comprise networks of components so-called systolic arrayswhich enable the simultaneous processing of huge amounts of knowledge. This efficient design saves power and computing time.

This is crucial because a single AI decision requires numerous calculations to be performed. Processors like Google's TPUs have gotten significantly higher on this regard in recent times.

In fact, the primary TPUs, first developed in 2015, were designed to hurry up the calculations performed by large, cloud-based servers throughout the training of AI models. In 2018, the primary TPU Designed to be used on computers on the “edge”, were released by Google. Then, in 2021, the primary TPUs designed for phones appeared – again for the Google Pixel.

There is great competition for the mixing of larger amounts of AI on mobile phones is growingThis implies that much more revolutionary technologies are prone to come onto the market in the approaching years.

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Must Read