Automated Inference on Criminality using Face ImagesXiaolin Wu, Xi Zhang
Publikationsdatum:
|
|
Diese Seite wurde seit 1 Jahr inhaltlich nicht mehr aktualisiert.
Unter Umständen ist sie nicht mehr aktuell.
Zusammenfassungen
In late 2016, engineering researchers Xiaolin Wu and Xi Zhang submitted an article titled “Automated Inference on Criminality Using Face Images” to a widely used online repository of research papers known as the arXiv. In their article, Wu and Zhang explore the use of machine learning to detect features of the human face that are associated with “criminality.” They claim that their algorithm can use simple headshots to distinguish criminals from noncriminals with high accuracy.
Von Xiaolin Wu, Xi Zhang im Text The Nature of Bullshit 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. Also, we find some discriminating structural features
for predicting criminality, such as lip curvature, eye inner
corner distance, and the so-called nose-mouth angle. Above
all, the most important discovery of this research is that
criminal and non-criminal face images populate two quite
distinctive manifolds. The variation among criminal faces
is significantly greater than that of the non-criminal faces.
The two manifolds consisting of criminal and non-criminal
faces appear to be concentric, with the non-criminal manifold lying in the kernel with a smaller span, exhibiting a
law of normality for faces of non-criminals. In other words,
the faces of general law-biding public have a greater degree of resemblance compared with the faces of criminals,
or criminals have a higher degree of dissimilarity in facial
appearance than normal people.
Von Xiaolin Wu, Xi Zhang im Text Automated Inference on Criminality using Face Images (2016) Bemerkungen
If this strikes you as frighteningly close to Philip K. Dick’s Precrime police in Minority Report, and to other dystopian science fiction, you’re not alone. The media thought so too. A number of technology-focused press outlets picked up on the story and explored the algorithm’s ethical implications. If an algorithm could really detect criminality from the structure of a person’s face, we would face an enormous ethical challenge. How would we have to adjust our notions of innocence and guilt if we could identify people as criminals even before they committed a crime?
Von Carl T. Bergstrom, Jevin D. West im Buch Calling Bullshit (2020) im Text The Nature of Bullshit We see two massive problems. The first is that the images of noncriminals were selected to cast the individuals in a positive light. By contrast, the images from the set of criminals are official ID photographs. While it is unclear exactly what this means, it’s safe to guess that these have been selected neither by the person depicted, nor with the aim of casting him in a favorable light. Thank goodness no one judges our characters based upon our driver’s license photos!
A second source of bias is that the authors are using photographs of convicted criminals. If there are facial differences between the two groups, we won’t know whether these differences are associated with committing crimes or with being convicted. Indeed, appearance seems to matter for convictions. A recent study reports that in the US, unattractive individuals are more likely to be found guilty in jury trials than their attractive peers.* Thus while the authors claim that their algorithm is free of human biases, it could be picking up on nothing but these biases.
Von Carl T. Bergstrom, Jevin D. West im Buch Calling Bullshit (2020) im Text The Nature of Bullshit A second source of bias is that the authors are using photographs of convicted criminals. If there are facial differences between the two groups, we won’t know whether these differences are associated with committing crimes or with being convicted. Indeed, appearance seems to matter for convictions. A recent study reports that in the US, unattractive individuals are more likely to be found guilty in jury trials than their attractive peers.* Thus while the authors claim that their algorithm is free of human biases, it could be picking up on nothing but these biases.
Dieser Text erwähnt ...
Begriffe KB IB clear | Kriminalität , machine learning |
Zitationsgraph
5 Erwähnungen
- New Dark Age - Technology and the End of the Future (James Bridle) (2018)
- Bitwise: A Life in Code (David Auerbach) (2018)
- Diskriminierungsrisiken durch Verwendung von Algorithmen - Eine Studie, erstellt mit einer Zuwendung der Antidiskriminierungsstelle des Bundes. (Carsten Orwat) (2019)
- Calling Bullshit - The Art of Skepticism in a Data-Driven World (Carl T. Bergstrom, Jevin D. West) (2020)
- The Atlas of AI (Kate Crawford) (2021)
Volltext dieses Dokuments
Automated Inference on Criminality using Face Images: Artikel als Volltext (: , 943 kByte; : ) |
Anderswo suchen
Beat und dieser Text
Beat hat Dieser Text während seiner Zeit am Institut für Medien und Schule (IMS) ins Biblionetz aufgenommen. Beat besitzt kein physisches, aber ein digitales Exemplar. Eine digitale Version ist auf dem Internet verfügbar (s.o.). Aufgrund der wenigen Einträge im Biblionetz scheint er es nicht wirklich gelesen zu haben. Es gibt bisher auch nur wenige Objekte im Biblionetz, die dieses Werk zitieren.