Gaussianization

By Scott Shaobing Chen, Ramesh A. Gopinath. 2000

163 citations recorded by the source corpus.

Subfields: Probabilistic & Bayesian Methods

How researchers see this paper

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

In Probabilistic & Bayesian Methods

Citations in the corpus

Density estimation using Real NVP (2016, methodology): “…a generative network g would be theoretically possible using the change of variable formula : This formula has been mentioned in several papers including the maximum likelihood formulation of independent components analysis (ICA) [4, 23], gaussianization [9] and deep density models [43, 14, 3].”
Masked Autoregressive Flow for Density Estimation (2017, methodology): “An early example is Gaussianization [4], which is based on successive application of independent component analysis.”
Sharp Minima Can Generalize For Deep Nets (2017, methodology): “For an invertible preprocessing ξ − 1 , e.g. feature standardization , whitening or gaussianization (Chen & Gopinath, 2001), we will call f ξ = f ◦ ξ the prediction function on the preprocessed input u = ξ − 1 ( x ) .”
Neural Spline Flows (2019, background): “Long-standing probabilistic models such as copulas [11] and Gaussianization [4] can simply represent complex marginal distributions that would require many layers of transformations in flow-based models like RealNVP and Glow.”
Iterative Gaussianization: From ICA to Random Rotations (2011, methodology): “This is consistent with previously reported results [17].”
Generative networks as inverse problems with Scattering transforms (2018, methodology): “In higher dimensions, this strategy has been extended by iteratively performing a Gaussianization of one-dimensional variables, through independent component analysis (Chen & Gopinath, 2000) or with random rotations (Laparra et al., 2011).”
Learning Intermediate-Level Representations of Form and Motion from Natural Movies (2012, methodology): “Note that transforming the marginal distributions of coefficients has been utilized by a number of image modeling researchers (Chen and Gopinath, 2000; Shan et al., 2007; Lyu and Simoncelli, 2009) and has an interesting relation to divisive normalization (Simoncelli, 1997; Schwartz and Simoncelli, 2”
A RAD approach to deep mixture models (2019, methodology): “We take a deeper look at the Gaussianization process involved in both models.”
Building Normalizing Flows with Stochastic Interpolants (2022, unknown): “Early works on exploiting transport maps for generative modeling go back at least to Chen & Gopinath (2000), which focuses on normalizing a dataset to infer its likelihood.”
Advances in Nonlinear Blind Source Separation (2003, methodology): “Chen and Gopinath propose an iterative Gaussianizat ion technique in [109] which could according to them provide a computationally efficient solution to nonlinear ICA.”
Neural Network Renormalization Group (2018, methodology): “Since the size of each block is independent of the system size in NeuralRG, one may even employ more general bijectors [77] com-”
Nonlinear Extraction of Independent Components of Natural Images Using Radial Gaussianization (2009, methodology): “After the linear ICA transformation has factorized the source density, one may apply a marginal gaussianization procedure, non-linearly mapping each component to have a unit-variance gaussian density using 1D histogram transforms (e.g., Chen & Gopinath, 2000).”