DL 14.3 Learning manifolds using autoencoders

This is a lecture on Learning manifolds using Autoencoders. The manifold hypothesis is that data concentrates around low-dimensional manifolds in ambient space. This lecture is about learning the structures of such manifolds.