Bad science, computational imperialism, and the economy of attention
January 19, 2021 9:34 PM Subscribe
Bad science, computational imperialism, and the economy of attention. Problems with the attempts to predict complex human traits like sexual and political orientations from facial features.
Phrenology is back! Where are my ceramic heads?
posted by q*ben at 9:43 PM on January 19, 2021 [2 favorites]
posted by q*ben at 9:43 PM on January 19, 2021 [2 favorites]
At this point, I pretty much have a knee-jerk reaction to anything that's "for profit" but there's so much terrible that comes from it. I have only relatively recently become aware of how publishing in the sciences and other areas have become such an exploitative process (other aspects of the sciences also exploit the researchers but that's a whole other discussion). But yeah, a lot of junk research gets published because of these models. And there's also the paywall to access good research (but that's a massive discussion in the lbirary world and kind of a derail, [fuck Elsevier]).
posted by NotTheRedBaron at 9:51 PM on January 19, 2021 [7 favorites]
posted by NotTheRedBaron at 9:51 PM on January 19, 2021 [7 favorites]
Where are my ceramic heads?
In the bathrooms. Where else?
posted by Splunge at 6:57 AM on January 20, 2021 [1 favorite]
In the bathrooms. Where else?
posted by Splunge at 6:57 AM on January 20, 2021 [1 favorite]
very cool to see the author of the trashfire "gaydar" paper coming back to make headlines again. Infering X from facial features is just a headling-generating machine, I guess? Just iterate through different Xs until you get a positive result and publish.
posted by BungaDunga at 10:53 AM on January 20, 2021 [2 favorites]
posted by BungaDunga at 10:53 AM on January 20, 2021 [2 favorites]
Another takedown of a silly "use" of science and technology (that is very similar to what we have here) is:
Can AI Tell Whether You're A Criminal From Your Face? | Machine Learning and Physiognomy
posted by RuvaBlue at 4:19 PM on January 20, 2021
Can AI Tell Whether You're A Criminal From Your Face? | Machine Learning and Physiognomy
posted by RuvaBlue at 4:19 PM on January 20, 2021
Brb implementing a facial recognition network that predicts whether an AI researcher is prone to promoting bad science. Next version will add grift-value and career-advancement predictions.
posted by runcifex at 9:16 AM on January 21, 2021
posted by runcifex at 9:16 AM on January 21, 2021
« Older What'd I miss? | Soon May the Wellerman Come Newer »
This thread has been archived and is closed to new comments
posted by mr_roboto at 9:38 PM on January 19, 2021