Substance over Style: Document-Level Targeted Content Transfer

TL;DR

This paper introduces a new task of document-level targeted content transfer, focusing on rewriting recipes to meet dietary constraints, and proposes a GPT-2 based model that outperforms existing methods in coherence, diversity, and constraint adherence.

Contribution

The paper presents a novel task of document-level targeted content transfer and a GPT-2 based model trained on aligned recipe pairs to address it.

Findings

Model outperforms existing methods in automatic and human evaluations.

Rewrites are coherent, diverse, and adhere to dietary constraints.

Analysis shows progress toward substantive constraint-based language generation.

Abstract

Existing language models excel at writing from scratch, but many real-world scenarios require rewriting an existing document to fit a set of constraints. Although sentence-level rewriting has been fairly well-studied, little work has addressed the challenge of rewriting an entire document coherently. In this work, we introduce the task of document-level targeted content transfer and address it in the recipe domain, with a recipe as the document and a dietary restriction (such as vegan or dairy-free) as the targeted constraint. We propose a novel model for this task based on the generative pre-trained language model (GPT-2) and train on a large number of roughly-aligned recipe pairs (https://github.com/microsoft/document-level-targeted-content-transfer). Both automatic and human evaluations show that our model out-performs existing methods by generating coherent and diverse rewrites that obey the constraint while remaining close to the original document. Finally, we analyze our model's rewrites to assess progress toward the goal of making language generation more attuned to constraints that are substantive rather than stylistic.