Facebook may be busy with virtual reality, chatbots, and 360-degree cameras, but the company hasn’t forgotten about your News Feed.
In fact, its latest work with artificial intelligence may make change the ever-scrolling landscape of our status updates and posts.
The tech company revealed concepts for how AI can augment its social media platform. Facebook’s Applied Machine Learning team has been working on overdrive – conducting nearly 50 times as many experiments on the site’s AI algorithm a day than compared to last year, producing some intriguing uses along the way.
What can all that processing power and boundless parameters do for your Facebook feed? Well, the company wagers that all that deep learning could come in handy in lowering language barriers, searching through image albums, or helping the visually-impaired “see” a photo.
Seamless searching & sharing
According to Facebook, half of all its users don’t speak English, and few are fluent in the language of another group of users. One aspect Facebook’s AI is working on is the ability to understand context from speech.
By being able to understand abbreviations, emoticons, or regional phrases from across the world, Facebook could instantly translate posts from friends you’ve met abroad – or international news sources – with better accuracy than an average copy-paste translation tool.
Another tool that would be of use to any fervent Facebook photographer is the ability to search albums by content. By typing “snow,” “lake,” or “suede jacket” into the search bar, the AI can actually detect images featuring those subjects, saving you from having to scroll through your post history for an hour just to show a friend that sweet jacket you took a picture of at a flea market last year.
An additional application of the AI’s object recognition would let Facebook name subjects in a photo out loud for those with sight impairment. By calling out objects in the forefront and background as they scroll around the image, an imaginary copy of the photo can be formed in a person’s mind to enjoy, as well as reinforce the AI’s ability to parse who or what is being tagged in a photo.
The AI’s ability to recognize subjects is also being applied to video, where a demonstration shows the machine sorting through fireworks, food, and – of course – cat videos without the need for humans to pre-tag them for content.
While the tech isn’t ready for the world at large yet, it probably won’t be long before we find ourselves logging onto a more intelligent Facebook…and we thought adding Vines as your profile picture was advanced.
via Blogger http://ift.tt/1oZssSZ