A New Learning Algorithm for Blind Signal Separation

By Shun-ichi Amari, Andrzej S. Cichocki, Howard H. Yang. 1995

How researchers see this paper

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

In Speech & Audio

Citations in the corpus

Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs) (2015, methodology): “The recently introduced Hessian-Free Optimization technique (Martens, 2010) and the Krylov Subspace Descent methods (Vinyals & Povey, 2012) use an extended Gauss-Newton approximation of the Hessian, therefore they can be interpreted as versions of natural gradient descent (Pascanu & Bengio, 2014).”
Deep learning in neural networks: An overview (2014, neutral): “Many UL methods are designed to maximize information-theoretic objectives (e.g., Linsker, 1988; Barlow et al., 1989; MacKay and Miller, 1990; Plumbley, 1991; Schmidhuber, 1992b,c; Schraudolph and Sejnowski, 1993; Redlich, 1993; Zemel, 1993; Zemel and Hinton, 1994; Field, 1994; Hinton et al., 1995; D”
Measuring and regularizing networks in function space (2018, background): “Amari originally developed the natural gradient in the light of information geometry and efﬁciency (Amari et al. (1996); Amari (1998)).”
ICA Using Spacings Estimates of Entropy (2003, methodology): “We then measured the " difference " between the true matrix A and the W returned by the algorithm, according to the well-known criterion (Amari error), due to Amari et al. [15].”
Joint Blind Source Separation by Multi-set Canonical Correlation Analysis (2009, methodology): “Therefore, we calculate ISI based on the average global matrix of all datasets.”
Blind Source Separation and Independent Component Analysis : A Review (2004, methodology): “Applying the natural gradient [5, 25], infomax learning rule at each frequency bin is”
Modelling matrix time series via a tensor CP-decomposition (2021, background): “When d̂ = d ≥ 2, L(A, Â) is the divergence measure of Amari et al. (1996). Denote by Id a d× d diagonal matrix with the main diagonal elements being either 1 or −1.”
Approach and applications of constrained ICA (2005, background): “The wide applicability of the ICA, including blind source separation [3]–[5], feature extraction [6], [7], and signal detection [8], etc., has resulted in many batch and adaptive algorithms [1], [9]–[13] for the efficient realization of the ICA.”
Self-Supervised Generation of Spatial Audio for 360 Video (2018, methodology): “While early methods present the problem as independent component analysis, and focused on maximizing the statistical independence of the extracted signals [23, 7, 6, 2], recent approaches focus on data-driven solutions.”
Blind Source Separation and Independent Component Analysis: A Review (2005, methodology): “Applying the natural gradient [5, 25], infomax learning rule at each frequency bin is”
Causality detection based on information-theoretic approaches in time series analysis (2007, methodology): “Using the truncated polynomial expansion for marginal pdf’s, Amari et al. [6] derived an approximation of the marginal entropy where k x 3 and k x 4 are third-and fourth-order cumulants.”
Regularization Networks and Support Vector Machines (2000, methodology): “Various methods for ICA have been developed in recent years [3,9,63,53,65].”
Advances in Nonlinear Blind Source Separation (2003, methodology): “Since (see [5] for a historical review and early references), several well-performing BSS and ICA algorithms [6], [7], [8], [9], [10], [ 11 ], [12] have been developed and applied to an increasing number of applications [13], [14], [15], [16], [17], [1], [18], [4].”
Information Theory, Inference, and Learning Algorithms (2004, unknown): “Further reading on blind separation, including non-ICA algorithms, can be found in (Jutten and Herault, 1991; Comon et al., 1991; Hendin et al., 1994; Amari et al., 1996; Hojen-Sorensen et al., 2002).”
On-line Variational Bayesian Learning (2003, unknown): “[1].”
Edgeworth Approximation of Multivariate Differential Entropy (2005, background): “…differential entropy (Shannon, 1948) based on a polynomial density expansion, the one-dimensional Edge-worth expansion, has become popular since it was introduced in independent component analysis (ICA) (Comon, 1994; Amari, Cichocki, & Yang, 1996) and projection pursuit (Jones & Sibson, 1987).”
Group information guided ICA for fMRI data analysis (2013, unknown): “These ICs can be estimated by optimizing their independence, such as minimizing their mutual information (Amari et al., 1996) or maximizing their non-Gaussianity measured by negentropy or kurtosis (Hyvarinen, 1999).”