Dense Distributions from Sparse Samples: Improved Gibbs Sampling Parameter Estimators for LDA

TL;DR

This paper presents a new method for estimating LDA parameters from Gibbs samples by leveraging full conditional distributions, improving accuracy with minimal additional computational cost, and outperforming traditional methods in experiments.

Contribution

It introduces a novel averaging technique for Gibbs sampling in LDA that combines advantages of CVB0 and CGS, applicable to existing implementations.

Findings

Consistent improvement over standard CGS in all tested conditions.

Outperforms CVB0 inference in most experimental scenarios.

Highlights the benefits of averaging over multiple samples for LDA estimation.

Abstract

We introduce a novel approach for estimating Latent Dirichlet Allocation (LDA) parameters from collapsed Gibbs samples (CGS), by leveraging the full conditional distributions over the latent variable assignments to efficiently average over multiple samples, for little more computational cost than drawing a single additional collapsed Gibbs sample. Our approach can be understood as adapting the soft clustering methodology of Collapsed Variational Bayes (CVB0) to CGS parameter estimation, in order to get the best of both techniques. Our estimators can straightforwardly be applied to the output of any existing implementation of CGS, including modern accelerated variants. We perform extensive empirical comparisons of our estimators with those of standard collapsed inference algorithms on real-world data for both unsupervised LDA and Prior-LDA, a supervised variant of LDA for multi-label classification. Our results show a consistent advantage of our approach over traditional CGS under all experimental conditions, and over CVB0 inference in the majority of conditions. More broadly, our results highlight the importance of averaging over multiple samples in LDA parameter estimation, and the use of efficient computational techniques to do so.