SciGuy42@alien.topBtoMachine Learning@academy.garden•[D] Why are ML model outputs not tested regarding statistical significance?English
1·
1 year agoI review for AAAI, NeuRIPS, etc. If a paper doesn’t report some notion of variance, standard deviation, etc., I have no choice but to reject since it’s impossible to tell whether the proposed approach is actually better. In the rebuttals, the author’s response is typically “well, everyone else also does it this way”. Ideally, I’d like to see an actual test of statistical significance.
“We need a strong leader…”
No, we don’t. Nobody has a real clue about how to get to AGI and there isn’t even a precise definition of AGI (my personal one is just by examples, namely R2D2, C3PO, or Commander Data but it’s just my personal one, not an objective definition, and even those aren’t “general” for 100% of problems but neither are humans).
There many individual leaders in AI and whether you choose one or more of them to focus on particular efforts, that’s up to you. Science is supposed to be democratic, not a dictatorship.