ICML 2015poster4 citations
Deterministic Independent Component Analysis
Ruitong Huang, Andras Gyorgy, Csaba Szepesvári
Abstract
We study independent component analysis with noisy observations. We present, for the first time in the literature, consistent, polynomial-time algorithms to recover non-Gaussian source signals and the mixing matrix with a reconstruction error that vanishes at a 1/\sqrtT rate using T observations and scales only polynomially with the natural parameters of the problem. Our algorithms and analysis also extend to deterministic source signals whose empirical distributions are approximately independent.
BibTeX
@InProceedings{pmlr-v37-huangb15,
title = {Deterministic Independent Component Analysis},
author = {Huang, Ruitong and Gyorgy, Andras and Szepesvári, Csaba},
booktitle = {Proceedings of the 32nd International Conference on Machine Learning},
pages = {2521--2530},
year = {2015},
editor = {Bach, Francis and Blei, David},
volume = {37},
series = {Proceedings of Machine Learning Research},
address = {Lille, France},
month = {07--09 Jul},
publisher = {PMLR},
pdf = {http://proceedings.mlr.press/v37/huangb15.pdf},
url = {https://proceedings.mlr.press/v37/huangb15.html},
abstract = {We study independent component analysis with noisy observations. We present, for the first time in the literature, consistent, polynomial-time algorithms to recover non-Gaussian source signals and the mixing matrix with a reconstruction error that vanishes at a 1/\sqrtT rate using T observations and scales only polynomially with the natural parameters of the problem. Our algorithms and analysis also extend to deterministic source signals whose empirical distributions are approximately independent.}
}