Abstract:Nonlinear independent component analysis (ICA) aims to uncover the true latent sources from their observable nonlinear mixtures. Despite its significance, the identifiability of nonlinear ICA is known to be impossible without additional assumptions. Recent advances have proposed conditions on the connective structure from sources to observed variables, known as Structural Sparsity, to achieve identifiability in an unsupervised manner. However, the sparsity constraint may not hold universally for all sources in practice. Furthermore, the assumptions of bijectivity of the mixing process and independence among all sources, which arise from the setting of ICA, may also be violated in many real-world scenarios. To address these limitations and generalize nonlinear ICA, we propose a set of new identifiability results in the general settings of undercompleteness, partial sparsity and source dependence, and flexible grouping structures. Specifically, we prove identifiability when there are more observed variables than sources (undercomplete), and when certain sparsity and/or source independence assumptions are not met for some changing sources. Moreover, we show that even in cases with flexible grouping structures (e.g., part of the sources can be divided into irreducible independent groups with various sizes), appropriate identifiability results can also be established. Theoretical claims are supported empirically on both synthetic and real-world datasets.

On Model Identifiability in Analytic Postnonlinear ICA

Identifiability of overcomplete independent component analysis

Generalizing Nonlinear ICA Beyond Structural Sparsity

On the Identifiability of Sparse ICA without Assuming Non-Gaussianity

The Incomplete Rosetta Stone Problem: Identifiability Results for Multi-View Nonlinear ICA

Blind Channel Identication of Nonlinear Folding Mixing Model

Fixed-point Blind Source Separation Algorithm Based on ICA

A New Method of Blind Source Separation Using Single-Channel ICA Based on Higher-Order Statistics

Blind Source Separation of Underwater Acoustic Signal by Use of Negentropy-Based Fast ICA Algorithm

Independent Component Analysis Over Galois Fields

Beyond Comon’s Identifiability Theorem for Independent Component Analysis

Blind Source Separation Based On Ica

Blind Non-Independent Image Separation Based on Independent Component Analysis

Study on the Indeterminacy of the Statistical Independence Based ICA Methods

An ICA Algorithm Based on Symmetric and Asymmetric Generalized Gaussian Model

Separation of Two Independent Sources by the Information-Theoretic Approach with Cubic Nonlinearity.

Overcomplete ICA with a Geometric Algorithm

Spatiotemporal Blind Source Separation Using Double-Sided Approximate Joint Diagonalization.

A Pseudo-Euclidean Iteration for Optimal Recovery in Noisy ICA

Uniqueness of linear independent component analysis and the Skitovitch-Darmois theorem

Generalizing Geometric ICA to Nonlinear Settings