Facebook emphasised that AI (artificial intelligence) is "central" to the company's various products. In a blog post, Facebook's Joaquin Quinonero Candela from Applied Machine Learning (AML) team detailed how AI is powering various Facebook experiences.
For auto-tagging on videos, Candela explained that the idea will be to search for people who featured in the video. As of now, Facebook has not revealed plans when the auto-tagging feature for videos will be available to general public. The feature will not only allow someone watching the video to find who featured in the video but also by tapping on the tag will be able to directly shift to the duration when the person appears in the video.
Candela also touched upon other services that will be powered by AI including real-time video classification which will allow Facebook to classify the video without using tags or surrounding content.
Also detailed was the Talking pictures feature for which Facebook is currently building systems that can understand images. The company calls it image segmentation, which allows Facebook to recognise individual objects in the image as well as their relation. "Using image segmentation we will be able to build more immersive experiences for the visually impaired with "talking images" you can read with your fingertips, as well as more powerful ways to search images. In one case here, we have the ability to search for "a photo of us five on skis on the snow, with a lake in the background and trees on both sides," explained Candela.