Data-to-Text Generation with Content Selection and Planning

TL;DR

This paper introduces a neural network model for data-to-text generation that explicitly incorporates content selection and planning, leading to improved performance over existing end-to-end models.

Contribution

It proposes a two-stage neural architecture that separates content planning from text generation, enhancing control and output quality in data-to-text tasks.

Findings

Outperforms strong baselines on RotoWire dataset

Improves state-of-the-art results

Both automatic and human evaluations confirm effectiveness

Abstract

Recent advances in data-to-text generation have led to the use of large-scale datasets and neural network models which are trained end-to-end, without explicitly modeling what to say and in what order. In this work, we present a neural network architecture which incorporates content selection and planning without sacrificing end-to-end training. We decompose the generation task into two stages. Given a corpus of data records (paired with descriptive documents), we first generate a content plan highlighting which information should be mentioned and in which order and then generate the document while taking the content plan into account. Automatic and human-based evaluation experiments show that our model outperforms strong baselines improving the state-of-the-art on the recently released RotoWire dataset.