See The World of AI From the Eyes of a Chip

Shaurya Sharma
3 min readDec 30, 2020

--

MediaTek is making strides in Artificial Intelligence like never before

What if your phone could see more than you can? What if your car could make life-saving decisions faster than you? What if your home devices could make you feel even more at home? All of this and more is possible with the help of Artificial Intelligence or AI as it is more commonly known. A majority of these applications can occur these days but they are situated in the cloud which isn’t exactly lightning speed.

Edge AI is a piece of technology developed by MediaTek, a leading SoC developing company, where they embed the AI tech directly inside your devices. This means you can access AI instantly without waiting, even when you’re not connected. It results in a faster response rate, greater privacy, and more functionality. Optimizing AI requires a whole new AI processing unit that is geared around neural network computing and at the same time can save up on energy consumption as compared to a CPU. This is exactly what the MediaTek Edge AI does so that you can do more for even longer.

AI is advancing at a rapid rate and is quickly reshaping the technology we use in our homes and outside our homes to change the way we interact. What MediaTek has created is an innovative heterogeneous computing architecture that directs the correct tasks to the right place in an instant, allowing you to get the best from each kind of processor. They call their new Edge AI platform the NeuroPilot, and it supports all of the most popular AI frameworks. MediaTek also provides the widest connectivity in the market which ensures that your devices always stay connected to leverage the power of the cloud instantly.

It’s interesting how Edge AI will integrate with our daily lives

Interestingly, MediaTek’s reach knows no bounds since their chipsets already feature in a clear majority of devices that we all use, right from smartphones to smart homes. MediaTek already powers voice assistants in smart speakers such as Amazon’s Alexa which supports ultra-low-power voice wake-up from standby, locally processed voiceprint recognition, voice commands, and more. This blend of Edge AI and the Internet of Things (IoT) takes intelligence and processing power to new levels. With their NeuroPilot Edge AI devices, they hope to enable the next generation of intelligent products that will be accessible to everyone.

How is this done? Data is localized in the device itself without the need to upload it to the cloud for processing. This helps in saving bandwidth, hiding latency, optimizing data security, and reducing packet loss during the transmission of data. Currently, the chip that features the most powerful Edge AI in a smartphone is the MediaTek Helio P90. It uses something called real-time object recognition, which basically translates to intelligently identifying a product in a photo, based on its analysis of the brand, colour, and size of the product in relation to the person in the photograph. This is one practical example of what the Edge AI can do. This kind of edge processing performance sorts through 200 images per second and is accurate up to 99%.

Be it inside your home with smart TVs, smart speakers or outside your home with your cars with their in-vehicle entertainment systems, the AI experience is all-permeating. The time is not far when we’ll just do away with those pesky always-disappearing remote controls and instead opt for facial recognition and voice control systems. Sounds straight out of Black Mirror, doesn’t it?

--

--

Shaurya Sharma
Shaurya Sharma

Written by Shaurya Sharma

Pop culture whiz. Social Media junkie. Web guru. Unapologetic Trash TV connoisseur. I write more than I read. Talk to me about all things Tech.

No responses yet