Quote note (#305)

Vindicating Lombroso:

We study, for the first time, automated inference on criminality based solely on still face images. Via supervised machine learning, we build four classifiers (logistic regression, KNN, SVM, CNN) using facial images of 1856 real persons controlled for race, gender, age and facial expressions, nearly half of whom were convicted criminals, for discriminating between criminals and non-criminals. All four classifiers perform consistently well and produce evidence for the validity of automated face-induced inference on criminality, despite the historical controversy surrounding the topic …

November 19, 2016admin 16 Comments »


16 Responses to this entry

  • Brett Stevens Says:

    Might be a good way to do an initial sieve for citizens to remove. Get rid of the 20% who commit 80% of the crime and things get a lot easier.


    G. Eiríksson Reply:

    It´s quite possible to calculate which belong to the class most likely to commit serious crimes. This has been possible for quite some time. People don´t want to learn such things about humanity, thus they don´t really want to solve the worst problems of humanity.

    In fact, even lying can be outdated.



    R. J. Moore II Reply:

    The urban poor and low-class ethnic minorities always commit the majority of crimes. Almost no one else would take the risk, they have alternatives. The exception is the state which is a criminal enterprise and ought to be violently suppressed. Cops are worse than gangsters.


    G. Eiríksson Reply:

    Cops are gangsters for the State. Gang is the prototypical male group. Even now in some fancily furnitured office, men gang upon men.

    Academics, schoolteachers I mean, are gangsters.

    Academics are intellectual slumlords.

    Or pseudo…

    Posted on November 19th, 2016 at 9:40 pm Reply | Quote
  • Waycis Says:

    So, we’re getting rid of Niggers, Beaners, and Mudslimes?


    G. Eiríksson Reply:

    >controlled for race, gender, age

    maybe u will be



    dvchydra Reply:

    Shanghai Jiao Tong University I expect the entire dataset was asian (han) males.


    Posted on November 19th, 2016 at 10:20 pm Reply | Quote
  • Luciferian Limerick Says:

    Shouldn’t someone tell the Chinese this type of research is not permitted?


    Posted on November 19th, 2016 at 11:22 pm Reply | Quote
  • John Hannon Says:

    Some fun to be had down the pub when the app comes out.


    John Hannon Reply:

    Maybe not so much fun using it to detect criminals at home however.
    Just imagine finding out you were married to one.


    Posted on November 20th, 2016 at 1:03 am Reply | Quote
  • dvchydra Says:

    They managed to get a CNN with 90% accuracy when identifying criminal Han Chinese males. The dataset is 1856 id photos. Their face analysis concludes that criminal males have a higher variance in appearance than non-criminal males.

    Next thing you know the Chinese citizenship score is going to include how criminal looking you are.


    Posted on November 20th, 2016 at 2:00 am Reply | Quote
  • SVErshov Says:

    similar stereotypes exist in most societies. go to some international airport and see who is getting out first on arrival. slim males with short hair cut, who fly business class, without much luggage. I’m sure 100% of them criminals. And who is not criminal? when in countries like France income tax is 60%. people become criminals by stupid laws, not neccessary by comitting crime intentionally. they should run some test on doctors and bankers, nearly 100% gauranted, same with MPs and presidents.


    G. Eiríksson Reply:

    I´d guess the test finds correlations using already caught criminals, which isn´t the same class as criminals that never get caught.


    Posted on November 20th, 2016 at 3:32 am Reply | Quote
  • Frog Do Says:

    Nope. Peer-review-science is just getting better at selling itself to different marketable sectors, testing the initial efforts in foreign markets as usual. It’s bullshit written in math. But being pandered to is always profitable for monkey handlers.


    Posted on November 20th, 2016 at 4:16 am Reply | Quote
  • GC Says:

    This would make one hell of an interesting mobile phone app.


    Posted on November 21st, 2016 at 10:33 am Reply | Quote
  • ILYA Says:

    Check pictures of politicians. Are they in a wrong business?

    Check pictures of mathematicians. Are they in a wrong business?

    Check pictures of your relatives. Are they in a wrong business?

    Check your selfies. Are you in a wrong business?


    Posted on November 30th, 2016 at 4:33 pm Reply | Quote

Leave a comment