The picture just screams "Insert penis here."
Posted on 01/14/2021 7:40:56 PM PST by algore
Ubiquitous facial recognition technology can expose individuals’ political orientation, as faces of liberals and conservatives consistently differ. A facial recognition algorithm was applied to naturalistic images of 1,085,795 individuals to predict their political orientation by comparing their similarity to faces of liberal and conservative others. Political orientation was correctly classified in 72% of liberal–conservative face pairs, remarkably better than chance (50%), human accuracy (55%), or one afforded by a 100-item personality questionnaire (66%). Accuracy was similar across countries (the U.S., Canada, and the UK), environments (Facebook and dating websites), and when comparing faces across samples. Accuracy remained high (69%) even when controlling for age, gender, and ethnicity. Given the widespread use of facial recognition, our findings have critical implications for the protection of privacy and civil liberties.
(Excerpt) Read more at nature.com ...
It’s all in the eyes
Of course, the Nazis had to make do with paper charts to classify undesirables. Facebook has helped automate the process.
skinny jeans = fag, or lib, or possible hispanic.
"...political orientation by facial recognition..."
This is silly, and, yes, ‘drivel’.
You can’t tell anything about any person simply by the way that they appear.
If you accept this, you are acquiescing to completely shallow - and largely FAKE - science; and your acceptance of it will come back to bite you in the ass.
They just proved there is a scientific basis for the accusations that liberals are “soy boys” and “ugly women”.
Gotta be the smirk.
[[Facial recognition technology can expose political orientation from naturalistic facial images]]
Yes, but the question on everyone’s mind is, can it identify what gender the person is?
The picture just screams "Insert penis here."
Exactly. I can spot soy boys, filthy hippies, and skater duuuuudes with 100% accuracy.
Hi JT,
First, thanks much for your cooking thread. You’ll find my limited participation there.
I’m speaking from a position of knowledge. I work in this field and have done significant research in AI. I hold active software patents.
It’s not a matter of “accepting” this. I’m simply reporting that AI can be trained in a variety of ways and then tested against real data. The research shows that some traits can be picked out from images of people with statistically valid results. It is about correlation, not causation. I cannot explain the results that make it so because AI is not based on algorithms. I can only observe the resulting correlations as shown by pure research.
It isn’t a matter of “biting me in the ass.” It really is about pure science and experimentation, not “belief.”
Of course, you can choose to believe anything you wish. I’m speaking about honest experimental results, double-blind tests.
With all due respect (and more, which I have for you), you may wish to open your mind and dig into this a bit. I think you might be surprised.
I was.
Yes, of course.
All AI training is based on very large data sets.
If the data set is small, the training is weaker.
Also, if the training data is faulty (e.g. images are not properly tagged), that will lead to questionable results.
A very real concern about AI is that bias will be introduced by the people processing the training data. For example, if the technicians that tag twitter accounts as “inducing violence” tag conservative accounts more than liberal accounts, the AI will learn that pattern.
The trainer’s intent may not be malicious, but the bias will, nonetheless, be induced into the AI. The bias is then propagated at scale.
I believe this is happening on all of the AI controlling the banning and censorship of accounts on social media.
Yes. Interesting how this isn’t too potentially dangerous to study, but racial signifiers and differences were deemed off-limits.
Lol!
If you can “pass” as a hippie, then you can be our embedded spy in the liberal swamp.
This sentence doesn't make sense.
One would of course expect that the accuracy would increase when adjusted for age, gender, etc.
Regards,
So, if the person had a Mao or Che poster in the background, that would help expose their political orientation?
Color me surprised!
Regards,
From the excerpt:
Accuracy remained high (69%) even when controlling for age, gender, and ethnicity.
As I noted above in an earlier posting, I find the word "even" here confusing.
Nevertheless, as the article explicitly points out, the study did take ethnicity into account.
Regards,
I’ve been called in for jury duty - Seattle area. I’m debating if wearing a white Tyvek suit, half-mask respirator and googles will get me off the jury or not. It might just be considered a plus by the prosecuting attorney. “Hey - here’s a person that follows all of our decrees!”
Okay - add a MAGA hat to the outfit!
If their studies are accurate, our brains don’t actually do it that well—not nearly as well as the technology.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.