Stochastic algorithms with descent guarantees for ICA
[edit]
Proceedings of Machine Learning Research, PMLR 89:15641573, 2019.
Abstract
Independent component analysis (ICA) is a widespread data exploration technique, where observed signals are modeled as linear mixtures of independent components. From a machine learning point of view, it amounts to a matrix factorization problem with a statistical independence criterion. Infomax is one of the most used ICA algorithms. It is based on a loss function which is a nonconvex loglikelihood. We develop a new majorizationminimization framework adapted to this loss function. We derive an online algorithm for the streaming setting, and an incremental algorithm for the finite sum setting, with the following benefits. First, unlike most algorithms found in the literature, the proposed methods do not rely on any critical hyperparameter like a step size, nor do they require a linesearch technique. Second, the algorithm for the finite sum setting, although stochastic, guarantees a decrease of the loss function at each iteration. Experiments demonstrate progress on the stateoftheart for large scale datasets, without the necessity for any manual parameter tuning.
Related Material


