Controllable Paraphrase Generation with a Syntactic Exemplar

TL;DR

This paper introduces a novel controllable paraphrase generation method that uses a syntactic exemplar to guide sentence syntax, employing a variational model with specialized modules and multitask training, validated on a new dataset.

Contribution

It proposes a new task of syntax control via exemplars, along with a variational model and a new dataset for evaluation, advancing controllable text generation techniques.

Findings

Model outperforms baselines in capturing syntax

Learned representations are more disentangled

Achieved improvements on a novel dataset

Abstract

Prior work on controllable text generation usually assumes that the controlled attribute can take on one of a small set of values known a priori. In this work, we propose a novel task, where the syntax of a generated sentence is controlled rather by a sentential exemplar. To evaluate quantitatively with standard metrics, we create a novel dataset with human annotations. We also develop a variational model with a neural module specifically designed for capturing syntactic knowledge and several multitask training objectives to promote disentangled representation learning. Empirically, the proposed model is observed to achieve improvements over baselines and learn to capture desirable characteristics.