Learning document embeddings along with their uncertainties

TL;DR

This paper introduces Bayesian SMM, a generative model that learns document embeddings as Gaussian distributions with uncertainty, improving data fit and robustness in topic identification over existing models.

Contribution

The paper presents Bayesian SMM, a novel generative model that encodes uncertainty in document embeddings and addresses intractability in variational inference.

Findings

Bayesian SMM outperforms neural variational models in perplexity on Fisher and 20Newsgroups datasets.

The model demonstrates robustness to overfitting in topic identification tasks.

Achieves comparable results to supervised models in unsupervised topic detection.

Abstract

Majority of the text modelling techniques yield only point-estimates of document embeddings and lack in capturing the uncertainty of the estimates. These uncertainties give a notion of how well the embeddings represent a document. We present Bayesian subspace multinomial model (Bayesian SMM), a generative log-linear model that learns to represent documents in the form of Gaussian distributions, thereby encoding the uncertainty in its co-variance. Additionally, in the proposed Bayesian SMM, we address a commonly encountered problem of intractability that appears during variational inference in mixed-logit models. We also present a generative Gaussian linear classifier for topic identification that exploits the uncertainty in document embeddings. Our intrinsic evaluation using perplexity measure shows that the proposed Bayesian SMM fits the data better as compared to the state-of-the-art neural variational document model on Fisher speech and 20Newsgroups text corpora. Our topic identification experiments show that the proposed systems are robust to over-fitting on unseen test data. The topic ID results show that the proposed model is outperforms state-of-the-art unsupervised topic models and achieve comparable results to the state-of-the-art fully supervised discriminative models.