Responses to Critiques on Machine Learning of Criminality Perceptions (Addendum of arXiv:1611.04135)

> The moment he stepped back on U.S. soil, he delivered a lecture titled “The Fallacy of Phrenology,” where he attacked the notion that head shape and size dictates the relative intelligence of different racial groups. But what if phrenology was based on machine learning? https://arxiv.org/abs/1611.04135v1 > 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. /s

boomark for next time a China hate-parade comes around: OP is treading the same water as https://arxiv.org/pdf/1611.04135v1.pdf "Automated Inference on Criminality using Face Images"

Explain the mechanism, please. It seems wildly implausible to me that someone's facial features change (let alone in a predictable way) once they've committed a crime. However, I can easily imagine this "working" by keying off things like age/race/gender, which will get you a value that's "better than chance" but isn't, really. Ditto for differences in the photos; nobody smiles in a mugshot, after all. (And yes, I know about the paper purporting to show this [https://arxiv.org/pdf/1611.04135.pdf] I just think it's ridiculous)

Oh, this has been done before: Automated Inference on Criminality using Face Images, Wu & Zhang, 2016, Arxiv. https://arxiv.org/pdf/1611.04135v2.pdf Deep Neural Networks Are More Accurate Than Humans at Detecting Sexual Orientation From Facial Images, Kosinski & Wang, Journal of Personality and Social Psychology 2018. https://www.gsb.stanford.edu/faculty-research/publications/d... Scientists Have Created an AI That Judges You by Your Looks, futurism.com. https://futurism.com/scientists-have-created-an-ai-that-judg... When artificial intelligence judges a beauty contest, white people win, qz.com. https://qz.com/774588/artificial-intelligence-judged-a-beaut...

And then there is obscene "research" like this: https://arxiv.org/pdf/1611.04135v1.pdf "Automated Inference on Criminality using Face Images." How does stuff like this get past IRB?

This article [0] discusses a paper by Chinese researchers [1] that did just that. (and yes, it is quite dumb/dangerous) [0] https://medium.com/@blaisea/physiognomys-new-clothes-f2d4b59... [1] https://arxiv.org/abs/1611.04135

Ah China and AI.. Just remembered this nice piece of paper: Automated Inference on Criminality using Face Images, https://arxiv.org/pdf/1611.04135v2.pdf At best they will become an Orwellian nightmare then they currently are, World domination? Please...

As a proof of concept, we have downloaded the pictures of 7000 corporate executives whose LinkedIn profiles suggest they work for financial organizations, and then averaged their faces to produce generalized white collar criminal subjects unique to each high risk zone. Future efforts will allow us to predict white collar criminality through real-time facial analysis. I'm guessing that's a riff on this? https://arxiv.org/pdf/1611.04135v1.pdf (Automated Inference on Criminality using Face Images) Too bad there haven't been enough arrests of financial criminals to get actual mugshots.

Here is a paper titled Automated Inference on Criminality using Face Images [0]. The system is trained on existing conviction records. Inheriting all the prejudices and biases of our current justice system. [0] https://arxiv.org/abs/1611.04135