advocates of #machinelearning, I am told that you all know that (current) #ML is limited. fair enough. but which limits are you willing to *publicly* acknowledge?https://twitter.com/NotSimplicio/status/1173373706674085888 …
-
https://thegradient.pub/nlps-clever-hans-moment-has-arrived/ … https://www.alexirpan.com/2018/02/14/rl-hard.html … https://link.medium.com/XNeEA1Yw1Z https://www.skynettoday.com/editorials/state_of_nmt … tbh, in my experience people who do research are usually the most aware and cynical about the methods' limitations (granted, a few high profile exceptions excepted).
they may be aware, but what point to do they make those limits clear to the public? or do they hide them from the public, leading to false expectations?
-
-
I mean, researchers are pretty busy people, most are preoccupied with doing the research and not communicating it to the public. And see above - plenty of blog posts making clear we are not as far along as it may seem, not to mention blog posts accompanying papers becoming norm.
-
Suggesting researchers "hide" limitations of ML from the public is a dramatic, tbh. My first hand impressions are that researchers tend to be cynical about how hard and brittle things are, and annoyed by media mischaracterization - eg people's reactions to OpenAI stuff, and like.
End of conversation
New conversation -
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.