Independent component analysis: algorithms and applications

By Aapo Hyvärinen, Erkki Oja. 2000

9,240 citations recorded by the source corpus.

Subfields: Analytical Chemistry, Artificial Intelligence, Speech & Audio, Signal Processing

How researchers see this paper

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

In Analytical Chemistry

In Artificial Intelligence

In Speech & Audio

In Signal Processing

Citations in the corpus

Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift (2015, background): “In contrast, W u + b is more likely to have a symmetric, non-sparse distribution, that is “more Gaus-sian” (Hyv¨arinen & Oja, 2000); normalizing it is likely to produce activations with a stable distribution.”
Shortcut learning in deep neural networks (2020, background): “The goal is to recover the true generating factors of the data distribution from observations [142] by identifying independent causal mechanisms [128].”
NICE: Non-linear Independent Components Estimation (2014, methodology): “Independent component analysis (ICA) (Hyvärinen and Oja, 2000), and more specifically its maximum likelihood formulation, learn an orthogonal transformation of the data, necessitating a costly orthogonalization proceedure between parameter updates.”
Diagnosing and Enhancing VAE Models (2019, methodology): “In this case, the model is identiﬁable up to an arbitrary permutation and scaling (Hyv¨arinen and Oja, 2000).”
Learning deep representations by mutual information estimation and maximization (2018, background): “Usefulness of a representation is not just a matter of information content: representational characteristics like independence also play an important role (Gretton et al., 2012; Hyv¨arinen & Oja, 2000; Hinton, 2002; Schmidhuber, 1992; Bengio et al., 2013; Thomas et al., 2017).”
Explaining Explanations: An Overview of Interpretability of Machine Learning (2018, methodology): “The problem of separating latent factors is an old problem that has previously been attacked using a variety of techniques such as Principal Component Analysis [61], Independent Component Analysis [62], and Nonnegative Matrix Factorization [63].”
Parseval Networks: Improving Robustness to Adversarial Examples (2017, methodology): “Also, note that quasi-orthogonalization procedures similar to the one described here have been successfully used previously in the context of learning overcomplete representations with independent component analysis (Hyvärinen & Oja, 2000).”
Flexibly Fair Representation Learning by Disentanglement (2019, background): “The search of independent latent components that explain observed data has long been a focus on the probabilistic modeling community (Comon, 1994; Hyvärinen & Oja, 2000; Bach & Jordan, 2002).”
On the Binding Problem in Artificial Neural Networks (2020, background): “Independent Component Analysis (ICA) models the observed data as linear combinations (mixtures) of unobserved random variables (sources) that are statistically independent (Hyvärinen and Oja, 2000).”
The Incomplete Rosetta Stone problem: Identifiability results for Multi-view Nonlinear ICA (2019, methodology): “This unsupervised learning method aims at providing a principled approach to disentanglement of independent latent components, blind source separation, and feature extraction [21].”
Interpretable Basis Decomposition for Visual Explanation (2018, methodology): “Principal Component Analysis [12] and Independent Component Analysis [11] have been widely used to disentangle a lowdimensional basis from high-dimensional data.”
Causal Interpretability for Machine Learning - Problems, Methods and Evaluation (2020, background): “earning is to break down the features into the independent latent variables that are highly correlated with meaningful patterns [29]. In traditional machine learning, approaches such as PCA [44], ICA [42] and spectrum analysis [100] are proposed to discover disentangled components of data. Recently,”
Single Channel Voice Separation for Unknown Number of Speakers Under Reverberant and Noisy Settings (2020, background): “Over the years, many attempts have been made to tackle this separation problem considering single microphone [1, 2], multiple microphones [3, 4], supervised and unsupervised learning [5, 6].”
Voice Separation with an Unknown Number of Multiple Speakers (2020, background): “The ability of humans to do so has inspired many computational attempts, with much of the earlier work focusing on multiple microphones and unsupervised learning, e.g., the independent component analysis approach (Hyvärinen & Oja, 2000).”
Manifold Gaussian Processes for regression (2014, background): “, 2008)), maximization of the variance (PCA (Pearson, 1901)), maximization of the statistical independence (ICA (Hyvärinen and Oja, 2000)), or the preservation of the distances between data (isomap (Tenenbaum et al.”
Normalization Propagation: A Parametric Technique for Removing Internal Covariate Shift in Deep Networks (2016, background): “As also claimed by Ioffe & Szegedy (2015); Hyv¨arinen & Oja (2000), we assume the pre-activation ( u ) has a Gaussian form.”
Tensor decompositions for learning latent variable models (2012, background): “Tensor decompositions have been used in signal processing and computational neuroscience for blind source separation and independent component analysis (ICA) [CJ10].”