Facebook Experiments With Machine Learning To Improve Its Facial and Gestures Detection Technology

Technology has progressed quite rapidly over the course of the past few years, but even now with our cutting edge tech some things are just a little too difficult for artificial intelligence (AI) to tackle which is a big part of the reason why Facebook has been working so hard to try and improve the machine learning processes that it takes part in.

One aspect of machine learning that a lot of people know about is deepfakes, and for the most part the reason why they know about it is because of the fact that we have all seen sinister videos in which deepfakes are used to try and create videos of people saying and doing things that they normally would not otherwise. This has some pretty scary implications for the world that we are living in, but at the same time Facebook is trying to manipulate the same kind of machine learning in an attempt to do good.

The experimentation that Facebook is doing is basically going to allow you to use this machine learning software to modify your appearance in pictures or video. While you will still look somewhat like yourself, there will be a few minor changes that are going to end up completely changing your appearance so much so that you will become pretty much unrecognizable, at least to facial recognition software that might be used to try and identify you.

We are living in a day and age where our privacy is compromised at every turn, and sometimes you might want to make a video or take part in a video call where you would not want to be recognized. At least having the option to make this a reality is going to be the sort of thing that gives a lot of people comfort as it will give them a little more control over how they appear.

Another thing that Facebook is applying AI to is gestures. Basically a machine learning software will be used to analyze how we gesture during normal conversations, and this will then be used to try and make virtual conversations more reminiscent of actual conversations that we might end up having in the real world. This type of machine learning can also be used by people who are doing animations since for the most part a lot of CGI depends on motion capture but this kind of machine learning may make motion capture obsolete and allow gestures to be put into a computer to be performed by a computer generated character.



Facebook is also working on using machine learning to try and give people automated fashion advice that is applicable in the real world.

Read next: Facebook is Using AI and Machine Learning to Protect Women Online
Previous Post Next Post