Gary Marcus

@Ylecun perpetuates false dichotomy here, between nature & nurture, without presenting data. Is there any evidence primates have fewer innate priors than mountain goats? Any proof that more innate priors actually entails less adaptivity? Better priors can yield *better* learning.https://twitter.com/ylecun/status/1180171721481113602 …

learning =/= adaptivity, though. Better priors, if known, can indeed confer better performance. At an extreme you hand code the solution directly and have zero free parameters. But the more priors/structure, the less a model can be applied to different tasks (adaptability)

So for a system that can solve a multitude of different tasks, the question is whether it is built of a multitude of specialized sub-components each with unique priors, or a less specialized mush, or something in between?

humans are most flexible learners; cognitive development literature and development neuroscience give strong evidence that they are endowed with many priors; see spelke’s work and my still relevant review in Birth of the mind.

What's the evidence that these are indeed priors in the true sense and not skills that are learned early in life? The literature shows these skills exist, it does not say how they came about. And also how do you disentangle them from any influence from evolution-based learning?

i just pointed you to a whole book that reviews this sort of thing, deprivation studies, experiments in developmental neuroscience, cognitive development etc. if you are genuinely interested read it.

Yes, thanks, but the question didn't stem from ignorance of the work, it was a genuinely questioning the conclusions that can be drawn from it. Can you point to specific papers (not books) that detail that these are in fact *not* skills learned early in life?