Adapting Language Models for Non-Parallel Author-Stylized Rewriting

TL;DR

This paper introduces a novel method for author-stylized rewriting using a fine-tuned language model with a denoising autoencoder loss, enabling style transfer without parallel data and improving stylistic alignment.

Contribution

It proposes a cascaded encoder-decoder framework with DAE loss for style adaptation, addressing the lack of parallel data in stylized text generation.

Findings

Outperforms state-of-the-art baselines in style alignment

Preserves original content effectively

Provides an interpretable evaluation framework

Abstract

Given the recent progress in language modeling using Transformer-based neural models and an active interest in generating stylized text, we present an approach to leverage the generalization capabilities of a language model to rewrite an input text in a target author's style. Our proposed approach adapts a pre-trained language model to generate author-stylized text by fine-tuning on the author-specific corpus using a denoising autoencoder (DAE) loss in a cascaded encoder-decoder framework. Optimizing over DAE loss allows our model to learn the nuances of an author's style without relying on parallel data, which has been a severe limitation of the previous related works in this space. To evaluate the efficacy of our approach, we propose a linguistically-motivated framework to quantify stylistic alignment of the generated text to the target author at lexical, syntactic and surface levels. The evaluation framework is both interpretable as it leads to several insights about the model, and self-contained as it does not rely on external classifiers, e.g. sentiment or formality classifiers. Qualitative and quantitative assessment indicates that the proposed approach rewrites the input text with better alignment to the target style while preserving the original content better than state-of-the-art baselines.