“Deep Learning: A Critical Appraisal”
▻https://arxiv.org/abs/1801.00631
A good overview of the limits of deep learning by Gary Marcus. Nothing new but it’s nice to have them in one place and with examples. Amongst them:
. Deep learning thus far is data hungry
. Deep learning thus far is shallow and has limited capacity for
transfer
. Deep learning thus far has no natural way to deal with
hierarchical structure
. Deep learning thus far is not sufficiently transparent
. Deep learning thus far is difficult to engineer with
(Here is a good intro to Hinton’s Capsule Networks: ▻https://www.youtube.com/watch?v=pPN8d0E3900
)