[Boards: 3 / a / aco / adv / an / asp / b / bant / biz / c / can / cgl / ck / cm / co / cock / d / diy / e / fa / fap / fit / fitlit / g / gd / gif / h / hc / his / hm / hr / i / ic / int / jp / k / lgbt / lit / m / mlp / mlpol / mo / mtv / mu / n / news / o / out / outsoc / p / po / pol / qa / qst / r / r9k / s / s4s / sci / soc / sp / spa / t / tg / toy / trash / trv / tv / u / v / vg / vint / vip / vp / vr / w / wg / wsg / wsr / x / y ] [Search | Free Show | Home]

>We study, for the first time, automated inference on cri

This is a blue board which means that it's for everybody (Safe For Work content only). If you see any adult content, please report it.

Thread replies: 22
Thread images: 5

File: c31c7c3331a623713fb82c04ecdffd2a.jpg (102KB, 736x1104px) Image search: [Google]
c31c7c3331a623713fb82c04ecdffd2a.jpg
102KB, 736x1104px
>We study, for the first time, automated inference on criminality based solely on still face images.

https://arxiv.org/pdf/1611.04135v1.pdf

Is this Cyberpunk?
>>
>>51755093
Incredibly
>>
>>51755093
No, it's phrenology.
>>
If an autocratic society could detect an inclination towards criminality with 80% accuracy, would plastic surgery be banned?
>>
File: wor.jpg (174KB, 1412x554px) Image search: [Google]
wor.jpg
174KB, 1412x554px
>>51755174

Call it what you want, but it appears to work.
>>
File: error.jpg (15KB, 335x99px) Image search: [Google]
error.jpg
15KB, 335x99px
>>51755197
No, it does not work. Pic related. For practical facial biometrics you need to have 5-nines reliability at a minimum, given the tens of millions of transactions you'll be scanning. This isn't even 1 nine reliable. And that's assuming a very homogenous Han population sample. This will fail in any practical application.

tl;dr it's shit science

Probably won't stop the CPP from adopting it though!
>>
File: Comrade.jpg (39KB, 600x615px) Image search: [Google]
Comrade.jpg
39KB, 600x615px
>>51755093
You can tell a lot about a man by how he looks, it's common knowledge. For fuck's sake, look at this comrade right here, isn't his personality obvious just from his face?
>>
>>51755244

What if you combined the face scanning with some other automated profiling method to increase reliability?
>>
File: lovecraft.jpg (59KB, 601x534px) Image search: [Google]
lovecraft.jpg
59KB, 601x534px
Lovecraft was right!
>>
>>51755197
>it appears to work, because an image on the Internet says so

This reminds me of something that happened to me yesterday. I was talking to my boss and mentioned the thing about the Macedonian teenagers getting rich making fake news clickbait. He didn't believe me, because he had never heard anything about it on CNN, MSNBC, etc.

I googled "macedo" and Google autocompleted "macedonia fake news", clicked the first link, and showed it to him. He did a 180 and immediately believed that it was true because I showed him the headline on a website. After a short discussion, he said "I can't believe people would be that easily fooled!" to which I pointed out that he didn't actually look at the URL and just assumed that because the website I showed him had the NBC logo, that it must be a legitimate website with factual information.
>>
>>51755395
>He didn't believe me, because he had never heard anything about it on CNN, MSNBC, etc.

Similarly, many will dismiss the claims of the Chinese face-scanning scientists outright unless a reputable Western authority grants its stamp.
>>
>>51755516
Or, you know, simple evidence >>51755244
>>
>>51755530

This is a first tentative step, and yet many would dismiss the entire field as tainted or misguided. Science!
>>
>>51755516

I don't mean to sound racist, but the quality of scientific research done in Asia is pretty fucking lousy. They have terribly low rates of falsifying hypotheses, some widespread cultural stigma against getting "the wrong answer" and proving something false, that's found across national lines far and wide. It's weird, really. (One of the smaller island nations is especially good, and a match for American and European scientists, I forget which one)
>>
>>51757114
>One of the smaller island nations is especially good, and a match for American and European scientists, I forget which one

Singapore?
>>
>>51757170

Possibly, but it seems like it was one of the lesser known ones. I'd have to dig up the numbers again. About four years back I read an overview of official publications in Asia and their rates of falsifying various "alernative medicine" claims, and compared to Western scientists, the numbers were crazy low. Throw 'em some hypothesis like "water has memory and it'll heal your soul if filled with positive emotions" and chances are they'll find it's plausible.
It's a great region if you're a quack with a lot of money who needs some "research" to back your bullshit.
>>
>>51755093
>>We study, for the first time, automated inference on criminality based solely on still face images.
I'm pretty sure skin-tone recognition has been around for years...
>>
>>51755552
>first tentative step.
>phrenology exists since the 18th century.
nigga, we got centuries of experience.
Also, you've far more reliable indicators of criminal tendencies, like social background, and yet we don't use it because that would be segregation, and we tend to limit ourselves to real infractions instead of future ones.

>>51755183
Proably not. In a cyberpunk setting, the poors wouldn't have access to it, and the exact items wouldn't be public, meaning that you have to disfigure yourself just to be sure, and Mr. Scarface will make people react negatively.
>>
>>51757261
I get you mate. Don't worry, damn SJWs don't bother us here.
>>
>>51755244
Otoh, you could very easily make a five nines phrenology-cam for terrorism or something.

> return (not_a_terrorist);
>>
>>51757249

Here in slavland we have homeopathy sections in state-sponsored clinics. I've never had any actual doctors I went to recommend anything from there, but the thing's MERE EXISTENCE shows that there's some levels of bullshit.

Also shit like posters that mention aural and prayer healing for anything, although thankfully as just a supplement to actual real medicines.
>>
>>51757292
>Proably not. In a cyberpunk setting, the poors wouldn't have access to it, and the exact items wouldn't be public, meaning that you have to disfigure yourself just to be sure, and Mr. Scarface will make people react negatively.
If we wait long enough, that guy who makes the threads of girls with lines painted on their faces will show up.
Thread posts: 22
Thread images: 5


[Boards: 3 / a / aco / adv / an / asp / b / bant / biz / c / can / cgl / ck / cm / co / cock / d / diy / e / fa / fap / fit / fitlit / g / gd / gif / h / hc / his / hm / hr / i / ic / int / jp / k / lgbt / lit / m / mlp / mlpol / mo / mtv / mu / n / news / o / out / outsoc / p / po / pol / qa / qst / r / r9k / s / s4s / sci / soc / sp / spa / t / tg / toy / trash / trv / tv / u / v / vg / vint / vip / vp / vr / w / wg / wsg / wsr / x / y] [Search | Top | Home]

I'm aware that Imgur.com will stop allowing adult images since 15th of May. I'm taking actions to backup as much data as possible.
Read more on this topic here - https://archived.moe/talk/thread/1694/


If you need a post removed click on it's [Report] button and follow the instruction.
DMCA Content Takedown via dmca.com
All images are hosted on imgur.com.
If you like this website please support us by donating with Bitcoins at 16mKtbZiwW52BLkibtCr8jUg2KVUMTxVQ5
All trademarks and copyrights on this page are owned by their respective parties.
Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.
This is a 4chan archive - all of the content originated from that site.
This means that RandomArchive shows their content, archived.
If you need information for a Poster - contact them.