Scalable MCMC for Mixed Membership Stochastic Blockmodels

Abstract

We propose a stochastic gradient Markov chain Monte Carlo (SG-MCMC) algorithm for scalable inference in mixed-membership stochastic blockmodels (MMSB). Our algorithm is based on the stochastic gradient Riemannian Langevin sampler and achieves both faster speed and higher accuracy at every iteration than the current state-of-the-art algorithm based on stochastic variational inference. In addition we develop an approximation that can handle models that entertain a very large number of communities. The experimental results show that SG-MCMC strictly dominates competing algorithms in all cases.

Cite

Text

Li et al. "Scalable MCMC for Mixed Membership Stochastic Blockmodels." International Conference on Artificial Intelligence and Statistics, 2016.

Markdown

[Li et al. "Scalable MCMC for Mixed Membership Stochastic Blockmodels." International Conference on Artificial Intelligence and Statistics, 2016.](https://mlanthology.org/aistats/2016/li2016aistats-scalable/)

BibTeX

@inproceedings{li2016aistats-scalable,
  title     = {{Scalable MCMC for Mixed Membership Stochastic Blockmodels}},
  author    = {Li, Wenzhe and Ahn, Sungjin and Welling, Max},
  booktitle = {International Conference on Artificial Intelligence and Statistics},
  year      = {2016},
  pages     = {723-731},
  url       = {https://mlanthology.org/aistats/2016/li2016aistats-scalable/}
}