Semi-Supervised Formality Style Transfer with Consistency Training

TL;DR

This paper introduces a semi-supervised method for formality style transfer that leverages source-side unlabeled data with consistency training, achieving state-of-the-art results with less than 40% of parallel data.

Contribution

It proposes a novel semi-supervised framework utilizing source-side unlabeled sentences and consistency training, improving formality transfer performance over previous cycle-reconstruction methods.

Findings

Achieves state-of-the-art results on GYAFC benchmark.

Effective data filtering strategies enhance model performance.

Performs well with significantly less parallel data.

Abstract

Formality style transfer (FST) is a task that involves paraphrasing an informal sentence into a formal one without altering its meaning. To address the data-scarcity problem of existing parallel datasets, previous studies tend to adopt a cycle-reconstruction scheme to utilize additional unlabeled data, where the FST model mainly benefits from target-side unlabeled sentences. In this work, we propose a simple yet effective semi-supervised framework to better utilize source-side unlabeled sentences based on consistency training. Specifically, our approach augments pseudo-parallel data obtained from a source-side informal sentence by enforcing the model to generate similar outputs for its perturbed version. Moreover, we empirically examined the effects of various data perturbation methods and propose effective data filtering strategies to improve our framework. Experimental results on the GYAFC benchmark demonstrate that our approach can achieve state-of-the-art results, even with less than 40% of the parallel data.