MedLDA: A General Framework of Maximum Margin Supervised Topic Models

TL;DR

MedLDA introduces a max-margin framework for supervised topic modeling that improves predictive performance over traditional likelihood-based methods, applicable to various types of data and models.

Contribution

The paper proposes MedLDA, a novel max-margin supervised topic model that enhances prediction accuracy and can be integrated with different topic modeling approaches.

Findings

MedLDA outperforms likelihood-based models on movie review data.

MedLDA achieves better classification accuracy on 20 Newsgroups.

Efficient variational methods enable scalable inference for MedLDA.

Abstract

Supervised topic models utilize document's side information for discovering predictive low dimensional representations of documents. Existing models apply the likelihood-based estimation. In this paper, we present a general framework of max-margin supervised topic models for both continuous and categorical response variables. Our approach, the maximum entropy discrimination latent Dirichlet allocation (MedLDA), utilizes the max-margin principle to train supervised topic models and estimate predictive topic representations that are arguably more suitable for prediction tasks. The general principle of MedLDA can be applied to perform joint max-margin learning and maximum likelihood estimation for arbitrary topic models, directed or undirected, and supervised or unsupervised, when the supervised side information is available. We develop efficient variational methods for posterior inference and parameter estimation, and demonstrate qualitatively and quantitatively the advantages of MedLDA over likelihood-based topic models on movie review and 20 Newsgroups data sets.