Big tech companies like Apple, Google and Facebook has developed machine learning to improve the performance of software. It is usually a very centralize process, companies collect data on how you use their apps and store them in one place, later they train their algorithms using this aggregated data. Users can now have improved performance of the app or software they are using; for instance; if you do lot of searching in Google search engine; after some time you will start getting the search patterns and search terms that you often use, this way you do not need to type in the search words over and over again.
This method of software improvement is quite effective, but it requires updating apps, sending receiving data which is time consuming, it also affects user’s privacy because all the data back and forth is saved on their servers.
THE NEW EXPERIMENT ON GOOGLE AI AND ALGORITHMS
Google is now experimenting to curb these problems to train algorithms; Google is currently experimenting with new techniques of training AI what it calls Federated learning.
As previous AI learning was based on centralized approach, where Google has to save data on its servers, now Google is trying to decentralize the whole process. It will teach its AI and algorithms directly on user’s device, your phone’s CPU will be used to teach Google’s AI.
Google is currently testing this new method on keyboard app, Gboard on android devices. When keyboard will show user suggested queries, the app will remember those terms and compare it with what they ignored. This information is then used to make changes in app algorithms directly on the phone. Google has sliced down its machine learning software TensorFlow in order to carry out this essential training.
As Google explains in a blog post, this approach has a number of benefits. It’s more private, as the data used to improve the app never leaves users’ device; and it delivers benefits immediately, as users don’t have to wait for Google to issue a new app update before they can start using their personalized algorithms. Google says the whole system has been streamlined to make sure it doesn’t interfere with your phone’s battery life or performance. The training process only takes place when your phone is “idle, plugged in, and on a free wireless connection.”