Artificial intelligence is becoming the norm these days ever since digital assistants like Siri and Cortana first broke into the mainstream market. Today, Facebook is bringing artificial intelligence into its image search system that is capable of identifying items in the photos even if they aren’t tagged as such.
This A.I. utilises Facebook’s Lumos, a platform that was created by Facebook for the purpose of training A.I.s to better understand images and videos. Facebook has also utilised other traditional methods of training A.I. such as deep learning and a neural network. The result of the training? The aforementioned A.I. implemented into the image search can now pick up on scenes, objects, animals, places, attractions and clothing items.
Once the new image search system is rolled out, Facebook users can now conduct image searches so long as they remember a part of the content found in the photo. Once example that Facebook gave was that a user entering “black shirt photo” into the search box will cause the system to throw out all of the user’s photos that feature a black shirt in them, even though they weren’t tagged as such.
Source: Facebook