Tuesday, May 8, 2018

New Machine Learning features coming to Gmail and Google Photos

Google I/O 2018 has officially kicked off, and it's pretty clear that their main, initial focus is to evolve, improve and involve their artificial intelligence and machine learning features to make users lives way easier. Google Photos, the Google Camera, and the Google Assistant in itself are amazing showcases of Google's machine learning and artificial intelligence in action, but Google is going all-in on AI all across their services. That's why additional AI features will now be coming to two of Google's most-used services: Gmail and Google Photos.

These features were announced during Google's I/O 2018 keynote. Over the next weeks, Gmail will be receiving a "Smart Compose" feature. In the same fashion as Smart Reply, which suggests small, brief responses to your emails based on their content, Smart Compose will give you suggestions, like greetings, closings, and whole sentences in between, while writing an email from scratch or replying to an email. When you see a suggestion you like, you just click on the tab button and Gmail will output it to your email.

Google Photos, on the other hand, will be receiving "one tap actions" for applying to your photos. As soon as you open a photo inside the app, it will suggest actions based on your photo's conditions. Based on machine learning, it'll show you the option to rotate, brighten, or tweak a relevant photo to look better. This same machine learning will also allow Google Photos to automatically color relevant black and white photos, automatically detecting the subject's photo and coloring their skin, clothing, and more. While the backdrop remains black and white, future updates will also allow you to color the background and all kinds of pictures.

The new AI-powered Google Photos features will begin rolling out today as soon as the I/O keynote is over, while Smart Compose will be rolling out to end users in the next couple of weeks.


Source: Google 1 Source: Google 2



from xda-developers https://ift.tt/2wpFGBx
via IFTTT