Intrinsic Dimensionality and Autoencoders
Francisco Chinesta, Elías Cueto, Victor Champaney, Chady Ghnatios, Amine Ammar, Nicolas Hascoët, David González, Icíar Alfaro, Daniele Di Lorenzo, Angelo Pasquale, Dominique Baillargeat · Studies in big data · 2025
As indicated, manifold learning aims at extracting the low-dimensional manifold in which high-dimensional data are embedded, reducing the dimensionality, and consequantly the number of coordinates required to express the data. However, in the techniques just described the inverse mapping, that is, the high-dimensional data reconstruction from the reduced data, is a tricky issue.