Spectral Methods for Correlated Topic Models

TL;DR

This paper introduces spectral methods for learning a broad class of correlated topic models based on Normalized Infinitely Divisible distributions, extending LDA to handle arbitrary topic correlations efficiently.

Contribution

It develops guaranteed spectral algorithms for NID-based topic models, enabling efficient learning with only third-order moments and a novel diagonalization technique.

Findings

Improved perplexity on New York Times dataset

Enhanced modeling of topic correlations

Efficient learning with low computational complexity

Abstract

In this paper, we propose guaranteed spectral methods for learning a broad range of topic models, which generalize the popular Latent Dirichlet Allocation (LDA). We overcome the limitation of LDA to incorporate arbitrary topic correlations, by assuming that the hidden topic proportions are drawn from a flexible class of Normalized Infinitely Divisible (NID) distributions. NID distributions are generated through the process of normalizing a family of independent Infinitely Divisible (ID) random variables. The Dirichlet distribution is a special case obtained by normalizing a set of Gamma random variables. We prove that this flexible topic model class can be learned via spectral methods using only moments up to the third order, with (low order) polynomial sample and computational complexity. The proof is based on a key new technique derived here that allows us to diagonalize the moments of the NID distribution through an efficient procedure that requires evaluating only univariate integrals, despite the fact that we are handling high dimensional multivariate moments. In order to assess the performance of our proposed Latent NID topic model, we use two real datasets of articles collected from New York Times and Pubmed. Our experiments yield improved perplexity on both datasets compared with the baseline.