Alphabet Inc (NASDAQ:GOOGL) and Movidius have partnered towards bringing Machine Learning out of the cloud in proximity with people. Google will now begin licensing Myriad 2 family VPU chips from Movidius. These chips include MA2150 and MA2450 chips and their accompanying SDKs. With these chips, Google will seek to develop next generation devices.
Machine learning has slowly become important to Google with Google photo search and speech recognition using them. Machine learning is an artificial intelligence, which gives computers the ability to learn, find patterns and make predictions. The only challenge with Machine learning is that it only takes place in the cloud. With this partnership, Google will use Movidius’ chips to bring Machine Learning out of the cloud.
Computer vision is simply the ability of computing devices to view, interpret and understand images both in context and symbolically. For instance, a computer can record images of dogs catalog them, reference them, recognize them and compare them. The catch here is that the computer will extrapolate beyond just comparing pixels and familiar feature to understanding the essence of a dog without having to perform a 1 to 1 comparison.
Movidius CEO Remi El-Ouazzane stated during an interview that the vision they have for the company is to make devices more autonomous. By collaborating with Google, they have a greater probablity to bring the technology close to people. He also went ahead to state that, they had not discussed which devices will be built using this technology.
This technology will particularly be interesting on a phone. A phone, for example, might not need to access the internet whenever you want to use voice assistant feature. This would ultimately reduce latency and cut down on data usage. El-Ouazzane stated that devices, which will use this technology, are not limited to phones only but would be embedded in lots of devices regardless of which operating systems they are using. He pointed out to Google’s set of devices, which are being developed and stated that this would not be possible without such collaboration.
Machine learning might also become useful in autonomous drones among other related technology, which requires intelligence to happen quickly.