Faster independent component analysis by preconditioning with Hessian approximations
arXiv:1706.08171 · doi:10.1109/TSP.2018.2844203
Abstract
Independent Component Analysis (ICA) is a technique for unsupervised exploration of multi-channel data that is widely used in observational sciences. In its classic form, ICA relies on modeling the data as linear mixtures of non-Gaussian independent sources. The maximization of the corresponding likelihood is a challenging problem if it has to be completed quickly and accurately on large sets of real data. We introduce the Preconditioned ICA for Real Data (Picard) algorithm, which is a relative L-BFGS algorithm preconditioned with sparse Hessian approximations. Extensive numerical comparisons to several algorithms of the same class demonstrate the superior performance of the proposed technique, especially on real data, for which the ICA model does not necessarily hold.
23 pages, 3 figures
References in corpus (3)
Cited by in corpus (13)
- Decoding EEG Rhythms During Action Observation, Motor Imagery, and Execution for Standing and Sitting
- Revealing Preference in Popular Music Through Familiarity and Brain Response
- Spectral independent component analysis with noise modeling for M/EEG source separation
- A structured L-BFGS method and its application to inverse problems
- Blind Determination of the Number of Sources Using Distance Correlation
- Modeling Shared Responses in Neuroimaging Studies through MultiView ICA
- Boosting Independent Component Analysis
- Orthogonal Extended Infomax Algorithm
- Second-order Approximation of Minimum Discrimination Information in Independent Component Analysis
- Using image-extracted features to determine heart rate and blink duration for driver sleepiness detection
- Shared Independent Component Analysis for Multi-Subject Neuroimaging
- Independent mechanism analysis, a new concept?
- Adaptive Multi-View ICA: Estimation of noise levels for optimal inference