Plan-then-Generate: Controlled Data-to-Text Generation via Planning

TL;DR

This paper introduces a Plan-then-Generate framework that enhances control over structure in neural data-to-text generation, leading to improved quality and diversity of generated texts.

Contribution

The novel PlanGen framework enables explicit control of both intra- and inter-sentence structures in neural data-to-text models, advancing controllability in the field.

Findings

Achieves better control over generated text structure.

Improves generation quality and diversity.

Outperforms previous state-of-the-art methods.

Abstract

Recent developments in neural networks have led to the advance in data-to-text generation. However, the lack of ability of neural models to control the structure of generated output can be limiting in certain real-world applications. In this study, we propose a novel Plan-then-Generate (PlanGen) framework to improve the controllability of neural data-to-text models. Extensive experiments and analyses are conducted on two benchmark datasets, ToTTo and WebNLG. The results show that our model is able to control both the intra-sentence and inter-sentence structure of the generated output. Furthermore, empirical comparisons against previous state-of-the-art methods show that our model improves the generation quality as well as the output diversity as judged by human and automatic evaluations.