Microsoft today announced (opens in new tab) that it has improved its facial recognition API to better recognize gender and darker skin tones, addressing two weak points in currently available recognition technologies. The improvements, Microsoft says, have led to a reduction in error rates for men and women with darker skin by up to 20 times, and overall error rates for women by nine times.
Currently, facial recognition tech tends to perform best on men with lighter skin and worst on women with darker skin. Microsoft pins this built-in bias on the data used to train the AI behind facial recognition systems. In order to improve its accuracy, the company needed to factor in a diversity of people, as well as other factors like hairstyle, jewelry, and eyewear.
In addressing the problem, Microsoft's team of researchers sought to focus on three key areas of improvement for its datasets: age, gender, and skin tones. Hanna Wallach, a researcher in Microsoft's New York Lab said:
Going forward, Microsoft is looking to bring what it has learned about the elimination of bias in its AI tools to its other services. The effort will span "from idea creation and data collection to model training, deployment and monitoring," Microsoft says. Wallach explains:
Windows Central Newsletter
Get the best of Windows Central in in your inbox, every day!
Microsoft created the most unbiased AI ever invented and her name was Tay.
At first, I thought this was a joke and laughed. Turns out it was a troll instead.
That was due to the trolls and like any other person she acted as any teen would as in reflecting the environment around them. In that case it's peer pressure, in Tay's case it was trolls messing with the repeat after me capabality.
"silence is the gift of God" .. so why don't you just shut the eff up? :|
It still has problems with my facial hair and glasses. To the point where it is not useable.
Thank you for signing up to Windows Central. You will receive a verification email shortly.
There was a problem. Please refresh the page and try again.