Nonlinear principal component analysis using autoassociative neural networks

By Mark A. Kramer. 1991

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Deep learning in neural networks: An overview (2014, background): “feature hierarchies fed into SL classiﬁers (Behnke, 1999, 2003a), the Self-Organising Tree Algorithm (SOTA) (Herrero et al., 2001), and nonlinear Autoencoders (AEs) with more than 3 (e.g., 5) layers (Kramer, 1991; Oja, 1991; DeMers and Cottrell, 1993). Such AE NNs (Rumelhart et al., 1986) can be tra”
A practical tutorial on autoencoders for nonlinear feature fusion: Taxonomy, models, software and guidelines (2018, background): “Although the term autoencoder is the most popular nowadays, they were also known as autoassociative neural networks [38], diabolo networks [39] and replicator neural networks [40].”
ACAE-REMIND for Online Continual Learning with Compressed Feature Replay (2021, methodology): “Auto-encoders [18] learn representations in an unsupervised way by encouraging the model to reconstruct the input data.”
Nonlinear ICA through low-complexity autoencoders (1999, methodology): “The results above give rise to a new method for source separation: simply train autoencoders (e.g., [2, 3 , 4, 5]) via FMS.”
Dimensionality reduction strategy based on auto-encoder (2015, methodology): “In 1991, Kramer [8] introduced how to conduct nonlinear principal component analysis using auto-associative neural networks.”
Generalized Out-of-Distribution Detection: A Survey (2021, unknown): “Techniques include autoencoder (AE) (Kramer, 1991) and variational autoencoder (VAE) (Kingma & Welling, 2013)-based models, generative adversarial networks (GANs) (Goodfellow et al., 2014), ﬂow-based models (Rezende & Mohamed, 2015; Kobyzev et al., 2020), and representation enhancement strategies.”
A Fast and Accurate Face Detector Based on Neural Networks (2001, methodology): “Using three hidden layers of nonlinear neurons, an autoassociative neural network is able to perform a nonlinear dimensionality reduction [ 23 ].”
Feature Extraction Through LOCOCODE (1999, background): “Nonlinear codes have been obtained by nonlinear bottleneck AAs with more than three layers (e.g., Kramer, 1991; Oja, 1991; DeMers & Cottrell, 1993).”
The Method of Proper Orthogonal Decomposition for Dynamical Characterization and Order Reduction of Mechanical Systems: An Overview (2005, methodology): “Specifically , it was shown in reference [ 87 ] that if the motion of a nonlinear system is a single and synchronous NNM , then the mode given by a NLPCA analysis coincides with this NNM .”
Discriminative Autoencoders for Small Targets Detection (2014, methodology): “Other approaches learn the manifold in a global way, such as the Maximum Variant Unfolding [25], or autoencoders [26].”
Linear Versus Nonlinear PCA for the Classification of Hyperspectral Data Based on the Extended Morphological Profiles (2012, methodology): “On the other hand, Nonlinear Principal Component Analysis (NLPCA), performed using Autoassociative Neural Networks (AANNs) [9], produces a limited set of components in which the information content tends to be uniformly distributed.”
Visual Memorability for Robotic Interestingness via Unsupervised Online Learning (2020, methodology): “ing the density ratio estimation algorithm with the HOG feature [7]. However, in practice the approach cannot adapt well to changing distributions. In the long-term stage, we introduce an autoencoder [23] for unsupervised learning, which has been widely used for feature extraction in many applicatio”
A Unifying Review of Deep and Shallow Anomaly Detection (2020, methodology): “They were originally introduced during the 1980s [376]–[379] primarily as methods to perform nonlinear dimensionality reduction [380], [381], yet they have also been studied early on for AD [351], [352].”