Controllable Text Generation via Probability Density Estimation in the Latent Space

TL;DR

This paper introduces a novel controllable text generation method using probability density estimation in the latent space with Normalizing Flows, achieving state-of-the-art results in attribute relevance and text quality.

Contribution

The work presents a new control framework employing invertible transformations for flexible and effective control in text generation, surpassing previous methods in quality and diversity.

Findings

Outperforms strong baselines in attribute relevance and text quality.

Achieves state-of-the-art performance in controllable text generation.

Demonstrates flexible control strength adjustment.

Abstract

Previous work on controllable text generation has explored the idea of control from the latent space, such as optimizing a representation with attribute-related classifiers or sampling a representation from relevant discrete samples. However, they are not effective enough in modeling both the latent space and the control, leaving controlled text with low quality and diversity. In this work, we propose a novel control framework using probability density estimation in the latent space. Our method utilizes an invertible transformation function, the Normalizing Flow, that maps the complex distributions in the latent space to simple Gaussian distributions in the prior space. Thus, we can perform sophisticated and flexible control in the prior space and feed the control effects back into the latent space owing to the one-one-mapping property of invertible transformations. Experiments on single-attribute controls and multi-attribute control reveal that our method outperforms several strong baselines on attribute relevance and text quality and achieves the SOTA. Further analysis of control strength adjustment demonstrates the flexibility of our control strategy.