Topic Compositional Neural Language Model

TL;DR

This paper introduces a novel neural language model that combines global semantic understanding via topics with local word order modeling, improving language modeling and topic coherence.

Contribution

The paper presents a Topic Compositional Neural Language Model that integrates neural topic modeling with a Mixture-of-Experts RNN approach for enhanced language understanding.

Findings

Outperforms pure RNN and topic-guided models on multiple corpora

Produces coherent topics and meaningful sentence generation

Efficient training via matrix factorization of RNN weights

Abstract

We propose a Topic Compositional Neural Language Model (TCNLM), a novel method designed to simultaneously capture both the global semantic meaning and the local word ordering structure in a document. The TCNLM learns the global semantic coherence of a document via a neural topic model, and the probability of each learned latent topic is further used to build a Mixture-of-Experts (MoE) language model, where each expert (corresponding to one topic) is a recurrent neural network (RNN) that accounts for learning the local structure of a word sequence. In order to train the MoE model efficiently, a matrix factorization method is applied, by extending each weight matrix of the RNN to be an ensemble of topic-dependent weight matrices. The degree to which each member of the ensemble is used is tied to the document-dependent probability of the corresponding topics. Experimental results on several corpora show that the proposed approach outperforms both a pure RNN-based model and other topic-guided language models. Further, our model yields sensible topics, and also has the capacity to generate meaningful sentences conditioned on given topics.