MIT Researchers Design 10 Times Faster Chip Than Mobile GPUs, Can Run Artificial Intelligence Algorithms Locally


As the time passes, technology continues to touch newer heights, especially in the internet of things and connected devices fields. Reports claim that researchers at MIT have designed a chip, which can implement neutral networks without any hassle.


It’s claimed to be 10 times faster and more efficient than a mobile GPU, and can power mobile devices with powerful artificial intelligence algorithms without cloud support. So far, AI supportive devices used to connect to cloud for uploading and downloading data, but with this chip there’s no need for cloud support.

Neutral networks are implemented with the help of graphical processing units. As this new chip is better than GPUs, scientists are confident that it will reduce the overall processing time in the future. The announcement was first made at the International Solid State Circuits Conference in San Francisco where a number of business leaders, connected device manufacturers and other companies operating in M2M fields were present.

Neutral networks were widespread in the Pre-1970 era, but as the time passed, they lost their place in the market. Things got back on track when they revived under the name ‘deep learning’ during the past decade. Applications like speech, recognition, face detection, etc. use deep learning technology at a wide-scale. With current ecosystem in place, users find neutral networks very complicated. MIT researchers hope that this new chip will make things easy and hassle-free in the future.

You can click here and read more about this new invention.


Please enter your comment!
Please enter your name here