Topographic Independent Component Analysis

By Aapo Hyvärinen, Juha Karhunen, Erkki Oja. 2001

Subfields: Representation & Self-Supervised Learning, 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 Representation & Self-Supervised Learning

In Signal Processing

Citations in the corpus

Deep learning in neural networks: An overview (2014, background): “Many do this to uncover and disentangle hidden underlying sources of signals (e.g., Jutten and Herault, 1991; Schuster, 1992; Andrade et al., 1993; Molgedey and Schuster, 1994; Comon, 1994; Cardoso, 1994; Bell and Sejnowski, 1995; Karhunen and Joutsensalo, 1995; Belouchrani et al., 1997; Hyvärinen e”
RUDDER: Return Decomposition for Delayed Rewards (2018, background): “are introduced for starting an episode. The sequence of differences is deﬁned as 0:T := ( s 1;a 1;s0;a0);:::;( s T 1;a T 1;s T;a T) : (A120) 31 We assume that the differences are mutually independent [60]: p(( s t 1;a t 1;s t;a t) j( s 1;a 1;s0;a0);:::;( s t 2;a t 2;s t 1;a t 1); (A121) ( s t;a t;s ”
On the Transfer of Disentangled Representations in Realistic Settings (2020, background): “Ideas related to disentangling the factors of variation date back to the non-linear ICA literature (Comon, 1994; Hyv¨arinen and Pajunen, 1999; Bach and Jordan, 2002; Jutten and Karhunen, 2003; Hyvarinen and Morioka, 2016; Hyvarinen et al., 2019; Gresele et al., 2019).”
Weakly-Supervised Disentanglement Without Compromises (2020, background): “It has roots in the independent component analysis (ICA) literature, where the goal is to unmix independent non-Gaussian sources of a d-dimensional signal (Comon, 1994).”
A Meta-Transfer Objective for Learning to Disentangle Causal Mechanisms (2019, methodology): “, 2016) and independent component analysis models (Hyvärinen et al., 2001; Hyvärinen et al., 2018) are built on this assumption.”
Detecting non-causal artifacts in multivariate linear regression models (2018, methodology): “Model for confounding with uncorrelated sources Our model for the influence of the high-dimensional common cause Z on both X and Y is inspired by Independent Component Analysis (ICA) (Hyvärinen et al., 2001).”
Causal Discovery from Temporally Aggregated Time Series (2017, methodology): “This can be achieved by the independent component analysis (ICA) procedure (Hyvärinen et al., 2001), and (I −A) can be estimated up to row scaling and permutation indeterminacies.”
A Kernel Independence Test for Random Processes (2014, neutral): “, 2012), and independent component analysis (Hyvärinen et al., 2001).”
Improved contact prediction in proteins: using pseudolikelihoods to infer Potts models. (2012, methodology): “[24] A.”
Density Modeling of Images using a Generalized Normalization Transformation (2015, neutral): “The topographic ICA model (TICA; Hyvärinen et al., 2001) and the model presented in Köster & Hyvärinen (2010) are generalizations of ISA that are related to our model, but have more constrained nonlinearities.”
Music Source Separation in the Waveform Domain (2019, methodology): “Independent component analysis [Hyvärinen et al., 2004] relies on independence assumptions and multiple microphones to separate the sources.”
On the Identifiability of the Post-Nonlinear Causal Model (2009, neutral): “For example, under the condition that causal relations are linear and acyclic, the non-Gaussianity of the disturbances could help find the whole causal model uniquely (Shimizu et al., 2006), by resorting to the independent component analysis (ICA) technique (Hyvärinen et al., 2001).”
Estimating Divergence Functionals and the Likelihood Ratio by Convex Risk Minimization (2008, neutral): “In statistical machine learning and signal processing, divergences between probability distributions are frequently exploited as metrics to be optimized, such as in independent component analysis [6] and decentralized detection [25].”
Multi-digit Number Recognition from Street View Imagery using Deep Convolutional Neural Networks (2013, methodology): “We do not use any whitening (Hyv¨arinen et al. , 2001), local contrast normalization (Sermanet et al. , 2012), etc.”
Nonlinear ICA Using Auxiliary Variables and Generalized Contrastive Learning (2018, neutral): “This is remarkable since these principles are well-known in linear ICA literature, but they have been considered as two distinct principles [19, 2, 3, 13].”