Add to ORKGCollapsed Gibbs sampling is a frequently applied method to approximate intractable inte-grals in probabilistic generative models such as latent Dirichlet allocation. This sampling method has however the crucial drawback of high computational complexity, which makes it limited applicable on large data sets. We propose a novel dynamic sampling strategy to significantly improve the efficiency of collapsed Gibbs sampling. The strategy is explored in terms of efficiency, convergence and perplexity. Besides, we present a straight-forward parallelization to further improve the efficiency. Finally, we underpin our proposed im-provements with a comparative study on different scale data sets
The implementation of collapsed Gibbs samplers for non-parametric Bayesian models is non-trivial, re...
We introduce a methodology for performing approximate computations in very complex probabilistic sy...
Latent Dirichlet allocation (LDA) is a Bayesian network that has recently gained much popularity in ...
In this paper, we investigate combining blocking and collapsing - two widely used strategies for imp...
The hierarchical Dirichlet process (HDP) is an intuitive and elegant technique to model data with la...
Probabilistic inference in many real-world problems requires graphical models with deterministic alg...
Topic models for text analysis are most commonly trained using either Gibbs sampling or variational ...
<div><p>The partially collapsed Gibbs (PCG) sampler offers a new strategy for improving the converge...
The recent emergence of Graphics Processing Units (GPUs) as general-purpose parallel computing devic...
Rao-Blackwellisation is a technique that provably improves the performance of Gibbs sampling by summ...
The Partially Collapsed Gibbs (PCG) sampler offers a new strategy for improving the con-vergence of ...
We have a probabilistic statistical model which is required to adapt in the light of observed cases...
Sampling inference methods are computationally difficult to scale for many mod-els in part because g...
Rao-Blackwellisation is a technique that provably improves the performance of Gibbs sampling by summ...
Inference is a central problem in probabilistic graphical models, and is often the main sub-step in ...
The implementation of collapsed Gibbs samplers for non-parametric Bayesian models is non-trivial, re...
We introduce a methodology for performing approximate computations in very complex probabilistic sy...
Latent Dirichlet allocation (LDA) is a Bayesian network that has recently gained much popularity in ...
In this paper, we investigate combining blocking and collapsing - two widely used strategies for imp...
The hierarchical Dirichlet process (HDP) is an intuitive and elegant technique to model data with la...
Probabilistic inference in many real-world problems requires graphical models with deterministic alg...
Topic models for text analysis are most commonly trained using either Gibbs sampling or variational ...
<div><p>The partially collapsed Gibbs (PCG) sampler offers a new strategy for improving the converge...
The recent emergence of Graphics Processing Units (GPUs) as general-purpose parallel computing devic...
Rao-Blackwellisation is a technique that provably improves the performance of Gibbs sampling by summ...
The Partially Collapsed Gibbs (PCG) sampler offers a new strategy for improving the con-vergence of ...
We have a probabilistic statistical model which is required to adapt in the light of observed cases...
Sampling inference methods are computationally difficult to scale for many mod-els in part because g...
Rao-Blackwellisation is a technique that provably improves the performance of Gibbs sampling by summ...
Inference is a central problem in probabilistic graphical models, and is often the main sub-step in ...
The implementation of collapsed Gibbs samplers for non-parametric Bayesian models is non-trivial, re...
We introduce a methodology for performing approximate computations in very complex probabilistic sy...
Latent Dirichlet allocation (LDA) is a Bayesian network that has recently gained much popularity in ...