SoylentNews is people
SoylentNews
from the that's-a-silly-name-for-an-AI dept.
Stanford University researchers have used software in an attempt to determine sexual orientation from photos:
"Deep neural networks are more accurate than humans at detecting sexual orientation from facial images" is the title of an article by Stanford University's Michal Kosinski and Yilun Wang, to be published in the Journal of Personality and Social Psychology. The abstract:
We show that faces contain much more information about sexual orientation than can be perceived and interpreted by the human brain. We used deep neural networks to extract features from 35,326 facial images. These features were entered into a logistic regression aimed at classifying sexual orientation. Given a single facial image, a classifier could correctly distinguish between gay and heterosexual men in 81% of cases, and in 74% of cases for women. Human judges achieved much lower accuracy: 61% for men and 54% for women. The accuracy of the algorithm increased to 91% and 83%, respectively, given five facial images per person. Facial features employed by the classifier included both fixed (e.g., nose shape) and transient facial features (e.g., grooming style).
Consistent with the prenatal hormone theory of sexual orientation, gay men and women tended to have gender-atypical facial morphology, expression, and grooming styles. Prediction models aimed at gender alone allowed for detecting gay males with 57% accuracy and gay females with 58% accuracy. Those findings advance our understanding of the origins of sexual orientation and the limits of human perception. Additionally, given that companies and governments are increasingly using computer vision algorithms to detect people's intimate traits, our findings expose a threat to the privacy and safety of gay men and women.
The images and the sexual orientation information were drawn from an online dating site. Note that the study was limited to white people from the United States, because of the relative lack of images of nonwhite gays and lesbians on the site.
Also at TechCrunch, The Advocate, and The Guardian.
(Score: 2) by looorg on Monday September 11 2017, @07:57PM (4 children)
Why would anyone care? Does it really matter why? When we know that they will. The why only matters if we think we can somehow change their decision which I doubt we can. Since fetal selection is already a thing, it might be some sort of defect such as downs syndrome or it might just be that the fetus has the wrong gender compared to their preference. Technology in that regard will just give people more options to make selections out of.
As noted most people want "the best", whatever that is, for their offspring. They want them to have a normal life. We know that being a homosexual or being abnormal in some other fashion isn't exactly going to be smooth sailing. They, the parents, might be okay with their offspring being gay but large parts of society won't. It will be a thing they might want to adjust. Just like they might sooner or later be able to design their babies in other fashion such as why would anyone want a fat baby, or a baby that will grow up bald, or stupid or any kind of disease or defect etc. Then there might be the completely egotistical reasons such as them one day wanting to have cute little grandchildren. It's not like gay people can't produce that but it might not be the same for them since it's not done in the "normal" way. Not to mention the factor of "what will other people think!". A lot of factors beyond just the baby clearly influence or could influence the decision.
So they won't even have to be religious nutters, they might just have a desire for making things easy for their offspring. It they are nutters they probably don't even have issue with aborting or tweaking the fetus. It's all down to God, if he didn't want them to abort defective fetuses he wouldn't have made humans invent the technology to do it. I'm sure they can rationalize all their decisions if they just try hard enough.
Overall tho the story is odd, first of all there isn't a 100% sure answer. So the gaydar clearly have wiggle room. There is a massive slant in the testing images, only white american people cause they couldn't find enough images of gay people of other nationalities and races. All images from dating sites. One could perhaps assume that people there are trying to make themselves attractive to the kind of people they want to hook up with. Which might just enforce certain stereotypes. But over all computers are slightly or somewhat better at guessing then humans, humans seemed to have it down to about 3/5 while the computer brought it up to 4/5. It's not really a massive endorsement of the technology to just be slightly better then a human at guessing. Not sure if humans became better at guessing if it was more images, the computer became about 10% better if it got multiple images.
I recon one could even gay-it-up for a picture if one wanted to. It seems proper grooming is a factor, some makeup, proper lightning and such you might look totally gay (or not gay) in the eyes for the machine if you just wanted to.
(Score: 2) by Grishnakh on Monday September 11 2017, @08:13PM (2 children)
They, the parents, might be okay with their offspring being gay but large parts of society won't.
Depends on the society. Large parts of urban American society really don't care, and in fact homosexuals are very well accepted there. (Of course, those places probably aren't producing many kids on their own...)
It's not like gay people can't produce that but it might not be the same for them since it's not done in the "normal" way.
From what I've read, it's not that unusual for homosexuals to have children in the "normal" way. Many times, they do want kids, so if they're female, well it isn't exactly hard to get a female pregnant if you can find a willing male.... And it's a whole lot cheaper than an artificial clinical approach. (Obviously, this doesn't work for gay men, who would have to adopt unless they can find a willing egg donor + surrogate.)
It will be a thing they might want to adjust. Just like they might sooner or later be able to design their babies in other fashion such as why would anyone want a fat baby, or a baby that will grow up bald, or stupid or any kind of disease or defect etc. Then there might be the completely egotistical reasons such as them one day wanting to have cute little grandchildren.
Now this part is a little worrying. What other factors are choosy parents going to select against? I foresee a future where everyone genetically selects their kids to be extroverted sports stars and socialites, and within a generation there's not enough competent people left to keep society's technology going, and disaster ensures.
(Score: 2) by Arik on Tuesday September 12 2017, @04:02AM (1 child)
Don't look now, but we're there already, without needing designer babies. Social programs and state schools did that job just fine all on their own.
If laughter is the best medicine, who are the best doctors?
(Score: 2) by Grishnakh on Tuesday September 12 2017, @03:02PM
Don't be ridiculous. We have plenty of people going into STEM programs (maybe not as many as some would like...), and the pace of software development, as one example, certainly isn't slowing down. And there's plenty of people going into trades work (again, not as many as some would like) to do the hands-on work needed to keep things running. So no, we're not there yet.
(Score: 2) by FatPhil on Tuesday September 12 2017, @11:08AM
Given that the dataset they were using seems to be exactly 50/50, that isn't slightly better, that's very much better.
A coin-toss would be 50% accurate, a human is only 10% better than coin toss, but a computer is 30% better than coin toss. Another way of seeing how different these are is to view their being correct in terms of fair odds - the human has 3:2-on odds (which should cause a number like 1.5 to appear in your mind), and the neural net 4:1-on odds (which should cause a number like 4 to appear, 4 being very different from 1.5).
Great minds discuss ideas; average minds discuss events; small minds discuss people; the smallest discuss themselves