When Artificial Intelligence Meets Accessibility

Wednesday, 4 November 2015

Facebook’s AI team has been doing significant improvements on their researches.

Inspired by biology, the AI network is basically a network that simulate neuron connections.

They teach this virtual mind by showing some pictures and giving some input on what it is. For instance, some pictures of dogs and stating the breed, the AI starts to learn the characteristics of that breed and then it can recognize the dog breed in another picture with the dog in a different position.

Their artificial intelligence is about to help people. Now their advances enable their AI to read the images for you. The algorithm scrapes a picture and say what it may contain. The biggest bottlenecks here are the reliability and the speed on recognizing the photo elements. But it is just the start…

Regarding accessibility, the reading of images can provided vision-impaired people to picture the picture in their minds. The girl in the video says that only three words help to flash out the details she can’t see. Check out the full video here.

They are moving towards their goal: Build intelligent machines to help people in their daily lives.

Let’s hope they stay on this way.

Swift Is Now Open Source »