2013
NeurIPS
NeurIPS 2013
Stochastic Gradient Riemannian Langevin Dynamics on the Probability Simplex
Abstract
In this paper we investigate the use of Langevin Monte Carlo methods on the probability simplex and propose a new method, Stochastic gradient Riemannian Langevin dynamics, which is simple to implement and can be applied online. We apply this method to latent Dirichlet allocation in an online setting, and demonstrate that it achieves substantial performance improvements to the state of the art online variational Bayesian methods.
🌉
Interdisciplinary Bridge
- Artificial Intelligence and Mathematics & Optimization
🧭
Keyword Pioneer
- langevin monte carlo
🐣
Hot Topic Early Bird
- markov chain monte carlo
Authors
Topics
Artificial Intelligence > Bayesian & Probabilistic > Bayesian Learning
Machine Learning > Optimization & Theory > Bayesian Inference
Deep Learning > Models > Generative Models
Deep Learning > Models > Variational Inference
Mathematics & Optimization > Optimization > Stochastic Methods
Machine Learning > Bayesian & Probabilistic > Probabilistic Modeling
Artificial Intelligence > Bayesian & Probabilistic > Bayesian Inference
Machine Learning > Bayesian & Probabilistic > Variational Inference
Machine Learning > Core Methods > Topic Modeling
Machine Learning > Bayesian & Probabilistic > Markov Chain Monte Carlo
Keywords
stochastic gradient
langevin dynamics
online learning
variational inference
bayesian inference
latent dirichlet allocation
markov chain monte carlo
stochastic gradient riemannian langevin dynamics
probability simplex
topic model
stochastic gradient langevin dynamics
langevin monte carlo
variational bayesian method