Statistical learning theory of deep neural networks for data on a low-dimensional manifold

Inverse problems in machine learning where data exhibit low-dimensional structure