Linear and Parallel Learning of Markov Random Fields
Abstract
We introduce a new embarrassingly parallel parameter learning algorithm for Markov random fields which is efficient for a large class of practical models. Our algorithm parallelizes naturally over cliques and, for graphs of bounded degree, its complexity is linear in the number of cliques. Unlike its competitors, our algorithm is fully parallel and for log-linear models it is also data efficient, requiring only the local sufficient statistics of the data to estimate parameters.
Cite
Text
Mizrahi et al. "Linear and Parallel Learning of Markov Random Fields." International Conference on Machine Learning, 2014.Markdown
[Mizrahi et al. "Linear and Parallel Learning of Markov Random Fields." International Conference on Machine Learning, 2014.](https://mlanthology.org/icml/2014/mizrahi2014icml-linear/)BibTeX
@inproceedings{mizrahi2014icml-linear,
title = {{Linear and Parallel Learning of Markov Random Fields}},
author = {Mizrahi, Yariv and Denil, Misha and De Freitas, Nando},
booktitle = {International Conference on Machine Learning},
year = {2014},
pages = {199-207},
volume = {32},
url = {https://mlanthology.org/icml/2014/mizrahi2014icml-linear/}
}