Addressing "Documentation Debt" in Machine Learning Research: A Retrospective Datasheet for BookCorpus

TL;DR

This paper creates a preliminary datasheet for the BookCorpus dataset to address documentation gaps, revealing copyright issues, duplicates, genre skews, and other deficiencies to improve transparency in ML dataset usage.

Contribution

It provides the first systematic datasheet for BookCorpus, highlighting key deficiencies and advocating for better dataset documentation in machine learning research.

Findings

BookCorpus likely violates copyright restrictions.

Contains thousands of duplicated books.

Exhibits significant genre skew.

Abstract

Recent literature has underscored the importance of dataset documentation work for machine learning, and part of this work involves addressing "documentation debt" for datasets that have been used widely but documented sparsely. This paper aims to help address documentation debt for BookCorpus, a popular text dataset for training large language models. Notably, researchers have used BookCorpus to train OpenAI's GPT-N models and Google's BERT models, even though little to no documentation exists about the dataset's motivation, composition, collection process, etc. We offer a preliminary datasheet that provides key context and information about BookCorpus, highlighting several notable deficiencies. In particular, we find evidence that (1) BookCorpus likely violates copyright restrictions for many books, (2) BookCorpus contains thousands of duplicated books, and (3) BookCorpus exhibits significant skews in genre representation. We also find hints of other potential deficiencies that call for future research, including problematic content, potential skews in religious representation, and lopsided author contributions. While more work remains, this initial effort to provide a datasheet for BookCorpus adds to growing literature that urges more careful and systematic documentation for machine learning datasets.