# Variational Cross-domain Natural Language Generation for Spoken Dialogue Systems

## TL;DR

This paper introduces a variational autoencoder-based approach to enhance the diversity and quality of cross-domain natural language generation in spoken dialogue systems, especially with limited training data.

## Contribution

It proposes a novel RNN-based generator using conditional variational autoencoders to improve diversity and performance in cross-domain NLG tasks.

## Key findings

- Outperforms traditional RNN generators in diversity and accuracy
- Generates more diverse sentences with limited training data
- Achieves better results in cross-domain NLG scenarios

## Abstract

Cross-domain natural language generation (NLG) is still a difficult task within spoken dialogue modelling. Given a semantic representation provided by the dialogue manager, the language generator should generate sentences that convey desired information. Traditional template-based generators can produce sentences with all necessary information, but these sentences are not sufficiently diverse. With RNN-based models, the diversity of the generated sentences can be high, however, in the process some information is lost. In this work, we improve an RNN-based generator by considering latent information at the sentence level during generation using the conditional variational autoencoder architecture. We demonstrate that our model outperforms the original RNN-based generator, while yielding highly diverse sentences. In addition, our model performs better when the training data is limited.