Inference on Criminality using Face Images

vespasian

Brown Belt
@Brown
Joined
Jul 31, 2014
Messages
3,607
Reaction score
0
Convict-spotting algorithm criticized

http://www.bbc.com/news/technology-38092196

_92644191_mediaitem92644188.jpg


The paper is available on arxiv: https://arxiv.org/pdf/1611.04135v2.pdf

An experiment to see whether computers can identify criminals based on their faces has been conducted in China. Researchers trained an algorithm using more than 1,500 photos of Chinese citizens, hundreds of them convicts. They said the program was then able to correctly identify criminals in further photos 89% of the time.

As expected, this study ignited a firestorm in the liberal academia

"This is an example of statistics-led research with no theoretical underpinning," said Prof McVie. "What would be the reason that somebody's face would lead them to be criminal or not? There is no theoretical reason that the way somebody looks should make them a criminal.

Note that the critics are not disputing the findings of the study. They're saying there is no theoretical reason why the results should hold. Does this mean the study is invalid? Apparently so, according to the liberal mindset. If data does not conform to prejudices, it should be ignored or disputed.

"What worries me the most is that we might be judging who is a criminal based on their looks. That sort of approach went badly wrong in our not-too distant history."

So are the scientific merits of the work that are judged, or is it the liberal dogma that is being offended? Fortunately for them, the Chinese aren't being neutered by political correctness. No wonder they have only contempt for the West.
 
Not sure how it could work with Asians or black people.
 
Last edited:
These sorts of things will eventually be used in conjunction with other data analytic algorithms to create risk profiles.

By analyzing phone records, facebook, internet search history, economic activity (purchases), travel, work history, and any other data collected over a lifetime, a psychological profile and pre-crime score can be calculated or flags can be set depending on circumstances.

Eventually supercomputers could predict someones behavior better then that person themselves.

Slippery slope, if I do say so.
 
These sorts of things will eventually be used in conjunction with other data analytic algorithms to create risk profiles.

By analyzing phone records, facebook, internet search history, economic activity (purchases), travel, work history, and any other data collected over a lifetime, a pre-crime score can be calculated or flags can be set depending on circumstances.

Eventually supercomputers could predict someones behavior better then that person themselves.

Slippery slope, if I do say so.

can one find scientific counter-arguments instead of political correct ones to such research? PC idiots insist on banning this type of inquiry instead of taking it seriously
 
can one find scientific counter-arguments instead of political correct ones to such research? PC idiots insist on banning this type of inquiry instead of taking it seriously

This sort of research is happening, regardless if it is talked about publicly.
 
This sort of research is happening, regardless if it is talked about publicly.

you mean the FBI etc are doign this type of research? I highly doubt Obama would allow it
 
you mean the FBI etc are doign this type of research? I highly doubt Obama would allow it

Military/intelligence connected research basically yes. Sort of like how the massive international spy grid (NSA, five eyes) was worked on outside of mainstream knowledge for so long.

I doubt a president has much control over those types of things, and probably wouldn't even be aware of all the various research projects going on.

This type of technology can be used for more than just finding criminals though. General population management more like.

It would be a pretty big political hurdle though to have such analytics be incorporated in convicting people or other direct actions.

In China it would be much easier to implement because people have less rights.
 
Last edited:
WTF ppl, how the hell am i the first to post this reference to Chinese ppl and taking a picture?
9c354f616d1195d7d4186ff62ca953ff.jpg
 
can one find scientific counter-arguments instead of political correct ones to such research? PC idiots insist on banning this type of inquiry instead of taking it seriously
The lack of an underlying theory is a scientific problem. Without it, all you have is data mining for correlation.
 
We already have this program, it is installed in all of our minds at birth.
 
The lack of an underlying theory is a scientific problem. Without it, all you have is data mining for correlation.

That's fine. Discovering strong correlations is definitely of practical interest
 
Could you imagine the shit-storm this would create if they started using it in a multi-cultural country?

MainHeaderImage.jpg
 
The lack of an underlying theory is a scientific problem. Without it, all you have is data mining for correlation.

Why not posit it's a self-fulfilling prophecy in that people (subconsciously) treat some folks as criminals (due to their faces) and that somehow leads one to behave in that fashion? I'm operating on two basic premises. The first is that your appearance influences how others treat you. The second is that the way a person is treated affects their behavior.
 
Why not posit it's a self-fulfilling prophecy in that people (subconsciously) treat some folks as criminals (due to their faces) and that somehow leads one to behave in that fashion? I'm operating on two basic premises. The first is that your appearance influences how others treat you. The second is that the way a person is treated affects their behavior.
I suppose you can. But I was not speculating about explanations. Instead, I was responding to the claim that 'there isn't an underlying theory' is a PC criticism, not a scientific criticism.
 
Back
Top