Hacker News Reader

Comment by pavement | Hacker News Reader

I think the signal that the AI bot has identified is very likely to be related to the effects of differing hormonal exchanges and releases over long periods of time.

We know that hormonal derivatives like steroids can influence appearance, and also that ambient secretions offer signaling at a distance among humans, in circumstances not limited to the synchronization of menstrual cycles, at a minimum.

Take these factors together, and exclusive close proximity to gender concentrations based on preference likely results in mild changes in appearance that automated analysis can determine as a non-zero signal, and hazard an informed guess with. People are not wrong in claiming that this is a form of stereotyping, and that sometimes the stereotype will be incorrect.

People may not like it, but just because you can't see it, doesn't mean it's not there.

View on HN

Or their dataset is a dating site and they are training a neural net to see features that are not there. 70-80% accuracy for a neural net is pretty bad if you have taken a neural networks course. The fact that Stanford students are publishing this is probably not because it's sound scientifically, but because ostentatiously riotous subject matter will increase the number of citations dramatically. Frankly, I don't think it makes sense to infer sexual orientation based on facial recognition, and I don't think it is sound science behind any of these findings. Yes, there may be features that imply correlation one way or the other, but the fact that these features would be "undetectable by the naked eye but still visually present" leads to all sorts of murky science that is "based" on neural networks. Also, this sort of "science" will no doubt be used wrongly to classify people who have not even investigated their own sexual leanings at all. People will be saying that their 7 year old has a propensity to be gay or straight based on the length of their nose? Come now, this is just balderdash and it's no surprise that it's contentious. Trying to provide some sort of defense for what is clearly just the Academia version of Clickbait is also a waste of time.

It's not clear to me -- your comment seems to contain conflicting statements -- do you think no correlation could exist between facial characteristics and sexuality? If you think there might be, do you think no program could ever detect it?

To answer your first question: Correlation could exist, but the idea that they are correlated implies that my physical features will change when my sexual preference changes. Ponder that for a minute.

Sexuality is not a boolean value, it's not a yes or no, it's not a this or that, even for people who are attracted to females they are not attracted to every female, and people who are attracted to males are not attracted to every male. It's such a silly thing to try and hone in on, because it removes the primary deciding component of human sexuality from the study: personal, intimate, real human preference. It's like attempting to discover one's favorite flavor of bagel shmear based on high school portraits.

You may find some surprising results, realize neural nets are trainable phenomena, and then run around giving everyone plastic surgery because more pronounced cheek bones tend to prefer Strawberry Shmear, and if peoples' cheek bones were adjusted their flavor preference would change.

>the idea that they are correlated implies that my physical features will change when my sexual preference change

No, it doesn't. It means sexual orientation and hormonal/chemical composition might be related, and hormonal/chemical composition might affect facial composition.

Everything you wrote afterwards is just an emotional outburst to an uncomfortable results from a study you obviously didn't bother reading.

"Sexual Orientation and Hormonal Composition" is a whole field of research on its own. Sexual Identity and Portraits we scraped off a dating site is too tenuous a leap for me to appreciate as any sort of "data" for serious science. If they had taken 5000 people, asked them to self-identify, and took a blood sample and measured various hormones, that would be a lot more satisfying.

Yes, I am emotionally bothered by this whole enterprise, because sexual orientation is not something you can understand through portraits, and this study makes the claim that 7/10 times a computer can do it. That's not the case.

Let's say there's a 50% chance that hormonal/chemical composition and sexual orientation are related, and let's say there's a 50% chance that hormonal composition affects facial features. That means that for the both of them to work out we have a 25% chance they are correlated. That's a pretty generous estimate, given that there is no data supporting the second half of that equation where "hormonal composition affects facial composition" as far as I'm aware.

The best we are doing with this dataset is inferring hormonal composition from what someone put down as their sexual preference. Or am I missing a large piece of the puzzle?

> To answer your first question: Correlation could exist, but the idea that they are correlated implies that my physical features will change when my sexual preference changes. Ponder that for a minute.

Why couldn't physical features change?

I believe hormone changes can affect appearances, so it doesn't seem beyond the realm of possibility.

> Sexuality is not a boolean value, it's not a yes or no, it's not a this or that

Did it occur to you that possibly the correlation with facial characteristics could be a matter of degree?

Consider the correlation between facial characteristics and racial background. I'm half European, half Asian. There's going to be a stronger correlation between people with near 100% European ancestry and "European facial characteristics" than people with 50% European ancestry, than people with 25% European ancestry, etc

My point is that Sexuality is not classifiable like Eye Color is classifiable, and we should not treat it as such.

Saying 70-80% is bad for a neural network is like saying that $1000 is too high a price without specifying what's for sale. The paper was peer-reviewed and accepted by a leading psychology journal. The lead author is an associate professor, not a student. At my first read, nothing unscientific stands out - claims that you don't think it's sound science, clickbait, and balderdash don't do anything to change the facts or advance the debate. That's not how science works.

Inner Preference versus External Appearance. If making a connection between these two things makes sense, then we should bust out the cranial measuring devices and get rid of college entrance exams.

Making a broad claim about the usual accuracy of deep nets is extremely misleading; the bounding factor is the task, not the model.

You don't even have to take issue with the modeling or any other part of the methodology to debunk this study; it's the interpretation that is plainly ridiculous. As you said, these are pictures from a dating site; awfully likely that people choose to present themselves differently based on who they want to attract. Also no appropriate statistical test applied to differences in faces picked up by the model. See this for more: http://callingbullshit.org/case_studies/case_study_ml_sexual...

> People will be saying that their 7 year old has a propensity to be gay or straight based on the length of their nose?

What's the big deal? We've already got people thinking their toddlers are transgendered because they picked up a doll one time...

Please stop using this site primarily for politics and ideology—especially inflammatory tangents.

https://news.ycombinator.com/newsguidelines.html

You should probably stop allowing stories where the discussion can be taken in that direction from being posted then.

A good accuracy for any solution is based far more on the problem than the model trying to solve it. 70-80% for one problem space might be horrible, but be a dream for others.

Yes but if evolution provided you with a 70-80% accurate way of deciding which berries were poisonous, or which potential mates were going to kill you instead of procreate with you, the neural net would be of rather negligible benefit.

This bot's sample is a dating website. I'm going to bet my socks on the fact that all it does is detect fashion styles and trends and associate them to groups. The "male homosexual" group is most likely filled with individuals dressed in a more fashinable style who are also well groomed (eyebrows, etc.)

I am certain that you could introduce a lot of errors by adding a bunch of metrosexual but straight individuals into the mix.

The paper states that they used an existing net, VGG-Face, to exact facial features: https://osf.io/fk3xr/ (page 13)

>VGG-Face aims at representing a given face as a vector of scores that are as unaffected as possible by facial expression, background, lighting, head orientation, image properties such as brightness or contrast, and other factors that can vary across different images of the same person.

All the heavy lifting for this paper was done by logistic regression on the facial features reported by VGG-Face, they didn't train a DNN specifically for identifying sexuality from image pixel values. The algo wouldn't have seen clothing at all.

Thank you for the link. My respect for the researchers has increased.

I still think that while the results are accurate(ish), they reveal fashion trends and not genetic changes. Can in point, the composite average of homosexual male has glasses. Facial features of homosexual are often gender atypical but that's trough a lot of effort to conform with fashion trends (eyebrow reshaping, smoother skin, "kinkier" smile, etc.). This is especially true on dating sites. The whole nature vs nurture thing.

While I agree that hormone levels can impact facial features, I have yet to see a strong study proving that hormones levels can be directly linked to sexual orientation or gender identity.

Reply on news.ycombinator.com