6 ms·
Facebook already has access to a larger repository of photos going back over a decade and all the years inbetween along with decent face recognition to create a
by flocial 8y ago
Facebook already has access to a larger repository of photos going back over a decade and all the years inbetween along with decent face recognition to create a much bigger dataset than resorting to a hashtag challenge. But I guess that wouldn't be newsworthy.
- JaleDarvis 8y agoAccess to what?
- berbec 8y agoa larger repository of photos going back over a decade
- corobo 8y agoThat is assuming it's Facebook training the algorithm
- reaperducer 8y agoFacebook denies involvement. But Facebook is no longer worthy of blind trust.
- BurningFrog 8y agoThis is a circular argument.
- acct1771 8y agoJust say what you're saying. It's Feds. And they're the biggest social influencers around.
- macintux 8y agoAs has been pointed out, people identifying themselves makes it easier to figure out which 10-year-old and current photos are actually: * themselves * 10 years old
- justapassenger 8y agoAnd of course no one would lie about it!
- universenz 8y agoJokes on you, Facebook is actually training their AI for deep fakes!
- macintux 8y agoThe percentage of people lying about it, at least before the stage where people start piling on with jokes, is probably much lower than the general noise in the data set if you were to try to find the photos you want without help.
- Lerc 8y agoThat too is data. Those outliers just get sent to the not-to-be-trusted watch list.
- reaperducer 8y agoFacebook already has access to a larger repository of photos going back over a decade and all the years inbetween along with decent face recognition to create a much bigger dataset than resorting to a hashtag challenge. But I guess that wouldn't be newsworthy. This was addressed in the article: In various versions of the meme, people were instructed to post their first profile picture alongside their current profile picture, or a picture from 10 years ago alongside their current profile picture. So, yes: These profile pictures exist, they’ve got upload time stamps, many people have a lot of them, and for the most part they’re publicly accessible. But let's play out this idea. Imagine that you wanted to train a facial recognition algorithm on age-related characteristics and, more specifically, on age progression (e.g., how people are likely to look as they get older). Ideally, you'd want a broad and rigorous dataset with lots of people's pictures. It would help if you knew they were taken a fixed number of years apart—say, 10 years. ... In other words, it would help if you had a clean, simple, helpfully labeled set of then-and-now photos.
- Kylekramer 8y agoIt tries and fails to address it. I could barely count the number of zeros I would have to put in front of the 1 in the percentage of Facebook's photo data this meme covers. And a good portion of their dataset will have EXIF timestamps. Training an algorithm on the meme only would be insane waste of their data set.
- DoctorOetker 8y agoperhaps they forgot to split a test set :) while I'm sure they already have a great dataset going back far, most of that dataset will be a small fragment of the world that already had internet 10 years ago, in the mean time a lot of people in other parts of the world have come online and may have old pictures of themselves
- cmmartin 8y agoThey don't need everyone to train their AI to detect aging. They don't even need a substantial sample size. All humans age in relatively similar ways.
- matchagaucho 8y agoIt seems, at least, Microsoft is way beyond needing a 10-year challenge training dataset. Their "how old" facial analysis app has been around a few years and is remarkably accurate. https://www.how-old.net/ https://www.how-old.net/
- 0xfffff 8y agodid you read the article? thats even addressed