AUC = 1
-
-
-
Can't have Type I/II errors if all your training data belong to the same class and also so does your entire domain of applicability.
Show this thread -
The miracle is someone got paid millions to write a piece of software that boils down to return True
Show this thread -
Pinging
@lizthegrey because this really needs to go into our talkShow this thread -
Like the work thay
@mathbabedotorg and@zeynep and others have done one algorithmic bias is super relevant and important but it would be *refreshing* at this point if those were the ethics issues we had to be facing.Show this thread
End of conversation
New conversation -
-
-
@sufsters when you’re a fasc white boy who puts ML on their resume and then gets the job bc of “connections “. This is going to be in an ethics textbook one day... -
Seriously, the ethics textbooks will have a conversion rate between how immoral a technical decision is and how much you can pay someone to completely disregard it.
-
Sounds like the beginnings of a good Ph.D. Thesis... Not normally into quantifying morality (in terms of utilitarianism), but it seems like it might be a decent candidate for a different paradigm based on social responsibility...
End of conversation
New conversation -
-
-
Much science. Very objectivity. Truly usefulness yes
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.