Sparse Parallel Training of Hierarchical Dirichlet Process Topic Models
Machine Learning
2020-10-07 v2 Computation and Language
Information Retrieval
Machine Learning
Abstract
To scale non-parametric extensions of probabilistic topic models such as Latent Dirichlet allocation to larger data sets, practitioners rely increasingly on parallel and distributed systems. In this work, we study data-parallel training for the hierarchical Dirichlet process (HDP) topic model. Based upon a representation of certain conditional distributions within an HDP, we propose a doubly sparse data-parallel sampler for the HDP topic model. This sampler utilizes all available sources of sparsity found in natural language - an important way to make computation efficient. We benchmark our method on a well-known corpus (PubMed) with 8m documents and 768m tokens, using a single multi-core machine in under four days.
Cite
@article{arxiv.1906.02416,
title = {Sparse Parallel Training of Hierarchical Dirichlet Process Topic Models},
author = {Alexander Terenin and Måns Magnusson and Leif Jonsson},
journal= {arXiv preprint arXiv:1906.02416},
year = {2020}
}