Gary Marcus

This point has always been known. But again, it is a limitation of supervised learning, not of the architecture (deep or not). Geoff Hinton's focus on unsupervised learning for the last 40 years (and me for the last 20) stems from this.

I agree it's a limitation of sup learning *as conceived today in ML*. But schools teach wetware pupils every day and that's sup learning, which implies to me that it's the choice of algorithms and how you put them together (architecture).

School learning is only partially supervised. Schooling teaches a tiny amount of high-level knowledge. The vast majority of our knowledge is acquired early and in a self-supervised manner. Without this background knowledge, schooling would be ineffective. See The Cake.

I think you just made the point @GaryMarcus is making. And mine. To me, the architecture of wetware is set by Evolution and that early learning. Semi-sup and sup learning thereafter is tuning. Much ML pre-supposes algos in the former and jumps to the latter.

Those of us interested in the fundamental questions view Evolution also as a learning process. Appealing to innateness is appealing to a prior learning process. That process involves interactions with the world and with other people, virtually all unsupervised.

and we wonder if your definition of learning then becomes so inclusive as to include literally everything. at which point it is meaningless.

Training and goal fitting are the wrong way to look at evolution (and learning is the wrong word for almost everything). Evolution is genetic (innate) conditioning from environmental feedback over long periods of time.

Learning is the best word for all phenomena w/ regards to 'adaptive'/regulatory behav change. Even those that can observed in phylogenesis and even earlier changes in evolution incl longtime anorganic structural change dependent on contextual factors. Your alternatives would be?