DGST: a Dual-Generator Network for Text Style Transfer

TL;DR

DGST introduces a dual-generator architecture for text style transfer that operates without discriminators or parallel data, achieving competitive results on Yelp and IMDb datasets.

Contribution

The paper presents a simple dual-generator model for style transfer that eliminates the need for discriminators and parallel corpora, simplifying the architecture.

Findings

Achieves competitive performance on Yelp and IMDb datasets.

Does not rely on discriminators or parallel data.

Outperforms some complex baseline models.

Abstract

We propose DGST, a novel and simple Dual-Generator network architecture for text Style Transfer. Our model employs two generators only, and does not rely on any discriminators or parallel corpus for training. Both quantitative and qualitative experiments on the Yelp and IMDb datasets show that our model gives competitive performance compared to several strong baselines with more complicated architecture designs.