Autoencoders reloaded

Abstract

In Bourlard and Kamp (Biol Cybern 59(4):291–294, 1998), it was theoretically proven that autoencoders (AE) with single hidden layer (previously called “auto-associative multilayer perceptrons”) were, in the best case, implementing singular value decomposition (SVD) Golub and Reinsch (Linear algebra, Singular value decomposition and least squares solutions, pp 134–151. Springer, 1971), equivalent to principal component analysis (PCA) Hotelling (Educ Psychol 24(6/7):417–441, 1993); Jolliffe (Principal component analysis, springer series in statistics, 2nd edn. Springer, New York ). That is, AE are able to derive the eigenvalues that represent the amount of variance covered by each component even with the presence of the nonlinear function (sigmoid-like, or any other nonlinear functions) present on their hidden units. Today, with the renewed interest in “deep neural networks” (DNN), multiple types of (deep) AE are being investigated as an alternative to manifold learning Cayton (Univ California San Diego Tech Rep 12(1–17):1, 2005) for conducting nonlinear feature extraction or fusion, each with its own specific (expected) properties. Many of those AE are currently being developed as powerful, nonlinear encoder–decoder models, or used to generate reduced and discriminant feature sets that are more amenable to different modeling and classification tasks. In this paper, we start by recalling and further clarifying the main conclusions of Bourlard and Kamp (Biol Cybern 59(4):291–294, 1998), supporting them by extensive empirical evidences, which were not possible to be provided previously (in 1988), due to the dataset and processing limitations. Upon full understanding of the underlying mechanisms, we show that it remains hard (although feasible) to go beyond the state-of-the-art PCA/SVD techniques for auto-association. Finally, we present a brief overview on different autoencoder models that are mainly in use today and discuss their rationale, relations and application areas.

Fish Species Classification using Optimized Deep Learning Model

“…An auto encoder decoder is a deep learning method that uses unsupervised learning to conduct encoding and decoding [31] [32]. Similar to an artificial neural network, it consists of input, hidden, and output layers [33].…”

Variational autoencoders to predict DNA-methylation age and provide biological insights in age-related health and disease

“…The underlying architecture consists of autoencoding layers which first encode (compress) the data in a lowerdimensional latent space after which the data is decoded (reconstructed) into its original dimension. Here, the aim of the encoder phase is not only to reduce the data dimensionality, but to compress the data by removing redundant information while at the same time keeping the most important information relevant for the research question in this reduced representation 37 . While a traditional AE minimizes the reconstruction error during training and results in a non-regularized latent space (the decoder cannot be used to generate valid input data from vectors sampled from the latent space), a VAE is stochastic and learns the parameters of the data distribution, i.e.…”