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 …
-
Most serious researchers will refrain from both stating unproven limits, and from promising success using the current set of methods.
serious researchers should acknowledge potential limits at the end of every paper; they also shouldn’t invite the inference that if they have solve one hard problem, they have solved them all. but examine closely the ML that gets cited the most: https://arxiv.org/abs/1801.05667