WikiTableT: A Large-Scale Data-to-Text Dataset for Generating Wikipedia Article Sections

TL;DR

WikiTableT is a large-scale dataset designed for data-to-text generation of Wikipedia sections, enabling research on long-form, multi-domain text generation with diverse metadata and challenging coherence and factuality issues.

Contribution

We introduce WikiTableT, a comprehensive dataset for Wikipedia section generation, and benchmark various strategies, highlighting challenges in coherence and factuality.

Findings

Best models generate fluent, high-quality texts

Models struggle with coherence and factuality

Dataset covers diverse topics and generation tasks

Abstract

Datasets for data-to-text generation typically focus either on multi-domain, single-sentence generation or on single-domain, long-form generation. In this work, we cast generating Wikipedia sections as a data-to-text generation task and create a large-scale dataset, WikiTableT, that pairs Wikipedia sections with their corresponding tabular data and various metadata. WikiTableT contains millions of instances, covering a broad range of topics, as well as a variety of flavors of generation tasks with different levels of flexibility. We benchmark several training and decoding strategies on WikiTableT. Our qualitative analysis shows that the best approaches can generate fluent and high quality texts but they struggle with coherence and factuality, showing the potential for our dataset to inspire future work on long-form generation.